Clinical Practice Guideline for Blood-based Circulating Tumor DNA Assays.

Circulating tumor DNA (ctDNA) has emerged as a promising tool for various clinical applications, including early diagnosis, therapeutic target identification, treatment response monitoring, prognosis evaluation, and minimal residual disease detection. Consequently, ctDNA assays have been incorporated into clinical practice. In this review, we offer an in-depth exploration of the clinical implementation of ctDNA assays. Notably, we examined existing evidence related to pre-analytical procedures, analytical components in current technologies, and result interpretation and reporting processes. The primary objective of this guidelines is to provide recommendations for the clinical utilization of ctDNA assays.

Differential Diagnosis of Pleural Effusion Using Machine Learning.

Rationale: Differential diagnosis of pleural effusion is challenging in clinical practice. Objectives: We aimed to develop a machine learning model to classify the five common causes of pleural effusions. Methods: This retrospective study collected 49 features from clinical information, blood, and pleural fluid of adult patients who underwent diagnostic thoracentesis between October 2013 and December 2018. Pleural effusions were classified into the following five categories: transudative, malignant, parapneumonic, tuberculous, and other. The performance of five different classifiers, including multinomial logistic regression, support vector machine, random forest, extreme gradient boosting, and light gradient boosting machine (LGB), was evaluated in terms of accuracy and area under the receiver operating characteristic curve through fivefold cross-validation. Hybrid feature selection was applied to determine the most relevant features for classifying pleural effusion. Results: We analyzed 2,253 patients (training set, n = 1,459; validation set, n = 365; extra-validation set, n = 429) and found that the LGB model achieved the best performance in both validation and extra-validation sets. After feature selection, the accuracy of the LGB model with the selected 18 features was equivalent to that with all 49 features (mean ± standard deviation): 0.818 ± 0.012 and 0.777 ± 0.007 in the validation and extra-validation sets, respectively. The model's mean area under the receiver operating characteristic curve was as high as 0.930 ± 0.042 and 0.916 ± 0.044 in the validation and extra-validation sets, respectively. In our model, pleural lactate dehydrogenase, protein, and adenosine deaminase levels were the most important factors for classifying pleural effusions. Conclusions: Our LGB model showed satisfactory performance for differential diagnosis of the common causes of pleural effusions. This model could provide clinicians with valuable information regarding the major differential diagnoses of pleural diseases.

Machine learning model for predicting immediate postoperative desaturation using spirometry signal data.

Postoperative desaturation is a common post-surgery pulmonary complication. The real-time prediction of postoperative desaturation can become a preventive measure, and real-time changes in spirometry data can provide valuable information on respiratory mechanics. However, there is a lack of related research, specifically on using spirometry signals as inputs to machine learning (ML) models. We developed an ML model and postoperative desaturation prediction index (DPI) by analyzing intraoperative spirometry signals in patients undergoing laparoscopic surgery. We analyzed spirometry data from patients who underwent laparoscopic, robot-assisted gynecologic, or urologic surgery, identifying postoperative desaturation as a peripheral arterial oxygen saturation level below 95%, despite facial oxygen mask usage. We fitted the ML model on two separate datasets collected during different periods. (Datasets A and B). Dataset A (Normal 133, Desaturation 74) was used for the entire experimental process, including ML model fitting, statistical analysis, and DPI determination. Dataset B (Normal 20, Desaturation 4) was only used for verify the ML model and DPI. Four feature categories-signal property, inter-/intra-position correlation, peak value/interval variability, and demographics-were incorporated into the ML models via filter and wrapper feature selection methods. In experiments, the ML model achieved an adequate predictive capacity for postoperative desaturation, and the performance of the DPI was unbiased.

Preliminary analysis of predicting the first recurrence in patients with neovascular age-related macular degeneration using deep learning.

BACKGROUND: To predict, using deep learning, the first recurrence in patients with neovascular age-related macular degeneration (nAMD) after three monthly loading injections of intravitreal anti-vascular endothelial growth factor (anti-VEGF). METHODS: Optical coherence tomography (OCT) images were obtained at baseline and after the loading phase. The first recurrence was defined as the initial appearance of a new retinal hemorrhage or intra/subretinal fluid accumulation after the initial resolution of exudative changes after three loading injections. Standard U-Net architecture was used to identify the three retinal fluid compartments, which include pigment epithelial detachment, subretinal fluid, and intraretinal fluid. To predict the first recurrence of nAMD, classification learning was conducted to determine whether the first recurrence occurred within three months after the loading phase. The recurrence classification architecture was built using ResNet50. The model with retinal regions of interest of the entire region and fluid region on OCT at baseline and after the loading phase is presented. RESULTS: A total of 1,444 eyes of 1,302 patients were included. The mean duration until the first recurrence after the loading phase was 8.20 ± 15.56 months. The recurrence classification system revealed that the model with the fluid region of OCT after the loading phase provided the highest classification performance, with an area under the receiver operating characteristic curve (AUC) of 0.725 ± 0.012. Heatmap analysis revealed that three pathological fluids, subsided choroidal neovascularization lesions, and hyperreflective foci were important areas for the first recurrence. CONCLUSIONS: The deep learning algorithm allowed for the prediction of the first recurrence for three months after the loading phase with adequate feasibility. An automated prediction system may assist in establishing patient-specific treatment plans and the provision of individualized medical care for patients with nAMD.

Reclassification of variants of tumor suppressor genes based on Sanger RNA sequencing without NMD inhibition.

Introduction: RNA sequence analysis can be effectively used to identify aberrant splicing, and tumor suppressor genes are adequate targets considering their loss-of-function mechanisms. Sanger sequencing is the simplest method for RNA sequence analysis; however, because of its insufficient sensitivity in cases with nonsense-mediated mRNA decay (NMD), the use of cultured specimens with NMD inhibition has been recommended, hindering its wide adoption. Method: The results of Sanger sequencing of peripheral blood RNA without NMD inhibition performed on potential splicing variants of tumor suppressor genes were retrospectively reviewed. For negative cases, in which no change was identified in the transcript, the possibility of false negativity caused by NMD was assessed through a review of the up-to-date literature. Results: Eleven potential splice variants of various tumor suppressor genes were reviewed. Six variants were classified as pathogenic or likely pathogenic based on the nullifying effect identified by Sanger RNA sequencing. Four variants remained as variants of uncertain significance because of identified in-frame changes or normal expression of both alleles. The result of one variant was suspected to be a false negative caused by NMD after reviewing a recent study that reported the same variant as causing a nullifying effect on the affected transcript. Conclusion: Although RNA changes found in the majority of cases were expected to undergo NMD by canonical rules, most cases (10/11) were interpretable by Sanger RNA sequencing without NMD inhibition due to incomplete NMD efficiency or allele-specific expression despite highly efficient NMD.

Analysis of MicroRNA Signature Differentially Expressed in Pancreatic Islet Cells Treated with Pancreatic Cancer-Derived Exosomes.

Since the majority of patients with pancreatic cancer (PC) develop insulin resistance and/or diabetes mellitus (DM) prior to PC diagnosis, PC-induced diabetes mellitus (PC-DM) has been a focus for a potential platform for PC detection. In previous studies, the PC-derived exosomes were shown to contain the mediators of PC-DM. In the present study, the response of normal pancreatic islet cells to the PC-derived exosomes was investigated to determine the potential biomarkers for PC-DM, and consequently, for PC. Specifically, changes in microRNA (miRNA) expression were evaluated. The miRNA specimens were prepared from the untreated islet cells as well as the islet cells treated with the PC-derived exosomes (from 50 patients) and the healthy-derived exosomes (from 50 individuals). The specimens were subjected to next-generation sequencing and bioinformatic analysis to determine the differentially expressed miRNAs (DEmiRNAs) only in the specimens treated with the PC-derived exosomes. Consequently, 24 candidate miRNA markers, including IRS1-modulating miRNAs such as hsa-miR-144-5p, hsa-miR-3148, and hsa-miR-3133, were proposed. The proposed miRNAs showed relevance to DM and/or insulin resistance in a literature review and pathway analysis, indicating a potential association with PC-DM. Due to the novel approach used in this study, additional evidence from future studies could corroborate the value of the miRNA markers discovered.

Effects of Unripe Black Raspberry Extract Supplementation on Male Climacteric Syndrome and Voiding Dysfunction: A Pilot, Randomized, Double-Blind, Placebo-Controlled Trial.

Male climacteric syndrome (MCS) is a medical condition that can affect middle-aged men whose testosterone levels begin to decline considerably. These symptoms may include fatigue, decreased libido, mood swings, and disturbed sleep. MCS can be managed with lifestyle modifications and testosterone replacement. However, testosterone therapy may cause number of side effects, including an increased risk of cardiovascular issues. This study aims to evaluate the efficacy and safety of unripe black raspberry extract (BRE) against MCS and voiding dysfunction in men with andropause symptoms. A total of 30 subjects were enrolled and randomly assigned to the BRE group (n = 15) or the placebo group (n = 15). Participants were supplemented with 4800 mg BRE or placebo twice daily for 12 weeks. The impact of BRE was assessed using the Aging Male's Symptoms (AMS scale), International Prostate Symptom Score (IPSS) and the IPSS quality of life index (IPSS-QoL). Additionally, male sex hormones, lipid profiles, and anthropometric indices were assessed 6 and 12 weeks after treatment. The AMS scores did not differ significantly between the two groups. In the BRE group, the total IPSS and IPSS-QoL scores decreased significantly after 12 weeks compared to baseline (p < 0.05), but there was no significant difference compared to the placebo group. However, a significant difference was observed in the IPSS voiding symptoms sub-score compared to the placebo group. Furthermore, LDL-C and TC levels were also significantly lower in the BRE group than in the placebo group (p < 0.05). Collectively, the study provides strong evidence supporting the safety of BRE as a functional food and its supplementation potentially enhances lipid metabolism and alleviates MCS and dysuria symptoms, limiting the development of BPH.

Investigation of optimal convolutional neural network conditions for thyroid ultrasound image analysis.

Neural network models have been used to analyze thyroid ultrasound (US) images and stratify malignancy risk of the thyroid nodules. We investigated the optimal neural network condition for thyroid US image analysis. We compared scratch and transfer learning models, performed stress tests in 10% increments, and compared the performance of three threshold values. All validation results indicated superiority of the transfer learning model over the scratch model. Stress test indicated that training the algorithm using 3902 images (70%) resulted in a performance which was similar to the full dataset (5575). Threshold 0.3 yielded high sensitivity (1% false negative) and low specificity (72% false positive), while 0.7 gave low sensitivity (22% false negative) and high specificity (23% false positive). Here we showed that transfer learning was more effective than scratch learning in terms of area under curve, sensitivity, specificity and negative/positive predictive value, that about 3900 images were minimally required to demonstrate an acceptable performance, and that algorithm performance can be customized according to the population characteristics by adjusting threshold value.

Fully automatic volume measurement of the adrenal gland on CT using deep learning to classify adrenal hyperplasia.

OBJECTIVES: To develop a fully automated deep learning model for adrenal segmentation and to evaluate its performance in classifying adrenal hyperplasia. METHODS: This retrospective study evaluated automated adrenal segmentation in 308 abdominal CT scans from 48 patients with adrenal hyperplasia and 260 patients with normal glands from 2010 to 2021 (mean age, 42 years; 156 women). The dataset was split into training, validation, and test sets at a ratio of 6:2:2. Contrast-enhanced CT images and manually drawn adrenal gland masks were used to develop a U-Net-based segmentation model. Predicted adrenal volumes were obtained by fivefold splitting of the dataset without overlapping the test set. Adrenal volumes and anthropometric parameters (height, weight, and sex) were utilized to develop an algorithm to classify adrenal hyperplasia, using multilayer perceptron, support vector classification, a random forest classifier, and a decision tree classifier. To measure the performance of the developed model, the dice coefficient and intraclass correlation coefficient (ICC) were used for segmentation, and area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity were used for classification. RESULTS: The model for segmenting adrenal glands achieved a Dice coefficient of 0.7009 for 308 cases and an ICC of 0.91 (95% CI, 0.90-0.93) for adrenal volume. The models for classifying hyperplasia had the following results: AUC, 0.98-0.99; accuracy, 0.948-0.961; sensitivity, 0.750-0.813; and specificity, 0.973-1.000. CONCLUSION: The proposed segmentation algorithm can accurately segment the adrenal glands on CT scans and may help clinicians identify possible cases of adrenal hyperplasia. KEY POINTS: • A deep learning segmentation method can accurately segment the adrenal gland, which is a small organ, on CT scans. • The machine learning algorithm to classify adrenal hyperplasia using adrenal volume and anthropometric parameters (height, weight, and sex) showed good performance. • The proposed segmentation algorithm may help clinicians identify possible cases of adrenal hyperplasia.

Diagnostic Assessment of Deep Learning Algorithms for Frozen Tissue Section Analysis in Women with Breast Cancer.

PURPOSE: Assessing the metastasis status of the sentinel lymph nodes (SLNs) for hematoxylin and eosin-stained frozen tissue sections by pathologists is an essential but tedious and time-consuming task that contributes to accurate breast cancer staging. This study aimed to review a challenge competition (HeLP 2019) for the development of automated solutions for classifying the metastasis status of breast cancer patients. Materials and Methods: A total of 524 digital slides were obtained from frozen SLN sections: 297 (56.7%) from Asan Medical Center (AMC) and 227 (43.4%) from Seoul National University Bundang Hospital (SNUBH), South Korea. The slides were divided into training, development, and validation sets, where the development set comprised slides from both institutions and training and validation set included slides from only AMC and SNUBH, respectively. The algorithms were assessed for area under the receiver operating characteristic curve (AUC) and measurement of the longest metastatic tumor diameter. The final total scores were calculated as the mean of the two metrics, and the three teams with AUC values greater than 0.500 were selected for review and analysis in this study. RESULTS: The top three teams showed AUC values of 0.891, 0.809, and 0.736 and major axis prediction scores of 0.525, 0.459, and 0.387 for the validation set. The major factor that lowered the diagnostic accuracy was micro-metastasis. CONCLUSION: In this challenge competition, accurate deep learning algorithms were developed that can be helpful for making a diagnosis on intraoperative SLN biopsy. The clinical utility of this approach was evaluated by including an external validation set from SNUBH.

Intronic LINE-1 insertion in SLCO1B3 as a highly prevalent cause of rotor syndrome in East Asian population.

Rotor syndrome is caused by digenic loss-of-function variants in SLCO1B1 and SLCO1B3 but only a few studies have reported co-occurring inactivating variants from both genes. A rotor syndrome-causing long interspersed element-1 (LINE-1) insertion in SLCO1B3 had been reported to be highly prevalent in the Japanese population but there has been no additional report. In spite of its known association with various human diseases, LINE-1 is hard to detect with current sequencing technologies. In this study, we aimed to devise a method to screen the LINE-1 insertion variant and investigate the frequency of this variant in various populations. A chimeric sequence, that was generated by concatenating the reference sequence at the junction and a part of inserted LINE-1 sequence, was searched from 725 raw sequencing data files. In cases containing the chimeric sequence, confirmatory long-range PCR and gap-PCR were performed. In total, 95 (13.1%) of 725 patients were positive for the chimeric sequence, and all were confirmed to have the SLCO1B3 LINE-1 insertion by PCR-based tests. The same chimeric sequence was searched from the 1000 Genomes Project data repository and the carrier frequency was remarkably high in the East Asian populations (10.1%), especially in Southern Han Chinese (18.5%), but almost absent in other populations. This SLCO1B3 LINE-1 insertion should be screened in a population-specific manner under suspicion of Rotor syndrome and the methods proposed in this study would enable this in a simple way.

Enhanced Long-Term Reliability of Seal DeltaSpot Welded Dissimilar Joint between 6061 Aluminum Alloy and Galvannealed Steel via Excimer Laser Irradiation.

Structural-adhesive-assisted DeltaSpot welding was used to improve the weldability and mechanical properties of dissimilar joints between 6061 aluminum alloy and galvannealed HSLA steel. Evaluation of the spot-weld-bonded surfaces from lap shear tests after long-term exposure to chloride and a humid atmosphere (5% NaCl, 35 °C) indicated that the long-term mechanical reliability of the dissimilar weld in a corrosive environment depends strongly on the adhesive-Al6061 alloy bond strength. Corrosive electrolyte infiltrated the epoxy-based adhesive/Al alloy interface, disrupting the chemical interactions and decreasing the adhesion via anodic undercutting of the Al alloy. Due to localized electrochemical galvanic reactions, the surrounding nugget matrix suffered accelerated anodic dissolution, resulting in an Al6061-T6 alloy plate with degraded adhesive strength and mechanical properties. KrF excimer laser irradiation of the Al alloy before adhesive bonding removed the weakly bonded native oxidic overlayers and altered the substrate topography. This afforded a low electrolyte permeability and prevented adhesive delamination, thereby enhancing the long-term stability of the chemical interactions between the adhesive and Al alloy substrate. The results demonstrate the application of excimer laser irradiation as a simple and environmentally friendly processing technology for robust adhesion and reliable bonding between 6061 aluminum alloy and galvannealed steel.

Automatic tip detection of surgical instruments in biportal endoscopic spine surgery.

BACKGROUND: Recent advances in robotics and deep learning can be used in endoscopic surgeries and can provide numerous advantages by freeing one of the surgeon's hands. This study aims to automatically detect the tip of the instrument, localize a point, and evaluate the detection accuracy in biportal endoscopic spine surgery (BESS). The tip detection could serve as a preliminary study for the development of vision intelligence in robotic endoscopy. METHODS: The dataset contains 2310 frames from 9 BESS videos with x and y coordinates of the tip annotated by an expert. We trained two state-of-the-art detectors, RetinaNet and YOLOv2, with bounding boxes centered around the tip annotations with specific margin sizes to determine the optimal margin size for detecting the tip of the instrument and localizing the point. We calculated the recall, precision, and F1-score with a fixed box size for both ground truth tip coordinates and predicted midpoints to compare the performance of the models trained with different margin size bounding boxes. RESULTS: For RetinaNet, a margin size of 150 pixels was optimal with a recall of 1.000, precision of 0.733, and F1-score of 0.846. For YOLOv2, a margin size of 150 pixels was optimal with a recall of 0.864, precision of 0.808, F1-score of 0.835. Also, the optimal margin size of 150 pixels of RetinaNet was used to cross-validate its overall robustness. The resulting mean recall, precision, and F1-score were 1.000 ± 0.000, 0.767 ± 0.033, and 0.868 ± 0.022, respectively. CONCLUSIONS: In this study, we evaluated an automatic tip detection method for surgical instruments in endoscopic surgery, compared two state-of-the-art detection algorithms, RetinaNet and YOLOv2, and validated the robustness with cross-validation. This method can be applied in different types of endoscopy tip detection.

Comparing intra-observer variation and external variations of a fully automated cephalometric analysis with a cascade convolutional neural net.

The quality of cephalometric analysis depends on the accuracy of the delineating landmarks in orthodontic and maxillofacial surgery. Due to the extensive number of landmarks, each analysis costs orthodontists considerable time per patient, leading to fatigue and inter- and intra-observer variabilities. Therefore, we proposed a fully automated cephalometry analysis with a cascade convolutional neural net (CNN). One thousand cephalometric x-ray images (2 k × 3 k) pixel were used. The dataset was split into training, validation, and test sets as 8:1:1. The 43 landmarks from each image were identified by an expert orthodontist. To evaluate intra-observer variabilities, 28 images from the dataset were randomly selected and measured again by the same orthodontist. To improve accuracy, a cascade CNN consisting of two steps was used for transfer learning. In the first step, the regions of interest (ROIs) were predicted by RetinaNet. In the second step, U-Net detected the precise landmarks in the ROIs. The average error of ROI detection alone was 1.55 ± 2.17 mm. The model with the cascade CNN showed an average error of 0.79 ± 0.91 mm (paired t-test, p = 0.0015). The orthodontist's average error of reproducibility was 0.80 ± 0.79 mm. An accurate and fully automated cephalometric analysis was successfully developed and evaluated.

SnackVar: An Open-Source Software for Sanger Sequencing Analysis Optimized for Clinical Use.

Despite the wide application of next-generation sequencing, Sanger sequencing still plays a necessary role in clinical laboratories. However, recent developments in the field of bioinformatics have focused mostly on next-generation sequencing, while tools for Sanger sequencing have shown little progress. In this study, SnackVar (https://github.com/Young-gonKim/SnackVar, last accessed June 22, 2020), a novel graphical user interface-based software for Sanger sequencing, was developed. All types of variants, including heterozygous insertion/deletion variants, can be identified by SnackVar with minimal user effort. The featured reference sequences of all of the genes are prestored in SnackVar, allowing for detected variants to be precisely described based on coding DNA references according to the nomenclature of the Human Genome Variation Society. Among 88 previously reported variants from four insertion/deletion-rich genes (BRCA1, APC, CALR, and CEBPA), the result of SnackVar agreed with reported results in 87 variants [98.9% (93.0%; 99.9%)]. The cause of one incorrect variant calling was proven to be erroneous base callings from poor-quality trace files. Compared with commercial software, SnackVar required less than one-half of the time taken for the analysis of a selected set of test cases. We expect SnackVar to be a cost-effective option for clinical laboratories performing Sanger sequencing.

Effectiveness of transfer learning for enhancing tumor classification with a convolutional neural network on frozen sections.

Fast and accurate confirmation of metastasis on the frozen tissue section of intraoperative sentinel lymph node biopsy is an essential tool for critical surgical decisions. However, accurate diagnosis by pathologists is difficult within the time limitations. Training a robust and accurate deep learning model is also difficult owing to the limited number of frozen datasets with high quality labels. To overcome these issues, we validated the effectiveness of transfer learning from CAMELYON16 to improve performance of the convolutional neural network (CNN)-based classification model on our frozen dataset (N = 297) from Asan Medical Center (AMC). Among the 297 whole slide images (WSIs), 157 and 40 WSIs were used to train deep learning models with different dataset ratios at 2, 4, 8, 20, 40, and 100%. The remaining, i.e., 100 WSIs, were used to validate model performance in terms of patch- and slide-level classification. An additional 228 WSIs from Seoul National University Bundang Hospital (SNUBH) were used as an external validation. Three initial weights, i.e., scratch-based (random initialization), ImageNet-based, and CAMELYON16-based models were used to validate their effectiveness in external validation. In the patch-level classification results on the AMC dataset, CAMELYON16-based models trained with a small dataset (up to 40%, i.e., 62 WSIs) showed a significantly higher area under the curve (AUC) of 0.929 than those of the scratch- and ImageNet-based models at 0.897 and 0.919, respectively, while CAMELYON16-based and ImageNet-based models trained with 100% of the training dataset showed comparable AUCs at 0.944 and 0.943, respectively. For the external validation, CAMELYON16-based models showed higher AUCs than those of the scratch- and ImageNet-based models. Model performance for slide feasibility of the transfer learning to enhance model performance was validated in the case of frozen section datasets with limited numbers.

Challenge for Diagnostic Assessment of Deep Learning Algorithm for Metastases Classification in Sentinel Lymph Nodes on Frozen Tissue Section Digital Slides in Women with Breast Cancer.

PURPOSE: Assessing the status of metastasis in sentinel lymph nodes (SLNs) by pathologists is an essential task for the accurate staging of breast cancer. However, histopathological evaluation of SLNs by a pathologist is not easy and is a tedious and time-consuming task. The purpose of this study is to review a challenge competition (HeLP 2018) to develop automated solutions for the classification of metastases in hematoxylin and eosin-stained frozen tissue sections of SLNs in breast cancer patients. MATERIALS AND METHODS: A total of 297 digital slides were obtained from frozen SLN sections, which include post-neoadjuvant cases (n = 144, 48.5%) in Asan Medical Center, South Korea. The slides were divided into training, development, and validation sets. All of the imaging datasets have been manually segmented by expert pathologists. A total of 10 participants were allowed to use the Kakao challenge platform for six weeks with two P40 GPUs. The algorithms were assessed in terms of the AUC (area under receiver operating characteristic curve). RESULTS: The top three teams showed 0.986, 0.985, and 0.945 AUCs for the development set and 0.805, 0.776, and 0.765 AUCs for the validation set. Micrometastatic tumors, neoadjuvant systemic therapy, invasive lobular carcinoma, and histologic grade 3 were associated with lower diagnostic accuracy. CONCLUSION: In a challenge competition, accurate deep learning algorithms have been developed, which can be helpful in making frozen diagnosis of intraoperative SLN biopsy. Whether this approach has clinical utility will require evaluation in a clinical setting.

Optimal matrix size of chest radiographs for computer-aided detection on lung nodule or mass with deep learning.

OBJECTIVES: To investigate the optimal input matrix size for deep learning-based computer-aided detection (CAD) of nodules and masses on chest radiographs. METHODS: We retrospectively collected 2088 abnormal (nodule/mass) and 352 normal chest radiographs from two institutions. Three thoracic radiologists drew 2758 abnormalities regions. A total of 1736 abnormal chest radiographs were used for training and tuning convolutional neural networks (CNNs). The remaining 352 abnormal and 352 normal chest radiographs were used as a test set. Two CNNs (Mask R-CNN and RetinaNet) were selected to validate the effects of the squared different matrix size of chest radiograph (256, 448, 896, 1344, and 1792). For comparison, figure of merit (FOM) of jackknife free-response receiver operating curve and sensitivity were obtained. RESULTS: In Mask R-CNN, matrix size 896 and 1344 achieved significantly higher FOM (0.869 and 0.856, respectively) for detecting abnormalities than 256, 448, and 1792 (0.667-0.820) (p < 0.05). In RetinaNet, matrix size 896 was significantly higher FOM (0.906) than others (0.329-0.832) (p < 0.05). For sensitivity of abnormalities, there was a tendency to increase sensitivity when lesion size increases. For small nodules (< 10 mm), the sensitivities were 0.418 and 0.409, whereas the sensitivities were 0.937 and 0.956 for masses. Matrix size 896 and 1344 in Mask R-CNN and matrix size 896 in RetinaNet showed significantly higher sensitivity than others (p < 0.05). CONCLUSIONS: Matrix size 896 had the highest performance for various sizes of abnormalities using different CNNs. The optimal matrix size of chest radiograph could improve CAD performance without additional training data. KEY POINTS: • Input matrix size significantly affected the performance of a deep learning-based CAD for detection of nodules or masses on chest radiographs. • The matrix size 896 showed the best performance in two different CNN detection models. • The optimal matrix size of chest radiographs could enhance CAD performance without additional training data.

Automated detection algorithm for C4d immunostaining showed comparable diagnostic performance to pathologists in renal allograft biopsy.

A deep learning-based image analysis could improve diagnostic accuracy and efficiency in pathology work. Recently, we proposed a deep learning-based detection algorithm for C4d immunostaining in renal allografts. The objective of this study is to assess the diagnostic performance of the algorithm by comparing pathologists' diagnoses and analyzing the associations of the algorithm with clinical data. C4d immunostaining slides of renal allografts were obtained from two different institutions (100 slides from the Asan Medical Center and 86 slides from the Seoul National University Hospital) and scanned using two different slide scanners. Three pathologists and the algorithm independently evaluated each slide according to the Banff 2017 criteria. Subsequently, they jointly reviewed the results for consensus scoring. The result of the algorithm was compared with that of each pathologist and the consensus diagnosis. Clinicopathological associations of the results of the algorithm with allograft survival, histologic evidence of microvascular inflammation, and serologic results for donor-specific antibodies were also analyzed. As a result, the reproducibility between the pathologists was fair to moderate (kappa 0.36-0.54), which is comparable to that between the algorithm and each pathologist (kappa 0.34-0.51). The C4d scores predicted by the algorithm achieved substantial concordance with the consensus diagnosis (kappa = 0.61), and they were significantly associated with remarkable microvascular inflammation (P = 0.001), higher detection rate of donor-specific antibody (P = 0.003), and shorter graft survival (P < 0.001). In conclusion, the deep learning-based C4d detection algorithm showed a diagnostic performance similar to that of the pathologists.