Performance changes due to differences among annotating radiologists for training data in computerized lesion detection

PURPOSE

The quality and bias of annotations by annotators (e.g., radiologists) affect the performance changes in computer-aided detection (CAD) software using machine learning. We hypothesized that the difference in the years of experience in image interpretation among radiologists contributes to annotation variability. In this study, we focused on how the performance of CAD software changes with retraining by incorporating cases annotated by radiologists with varying experience.

METHODS

We used two types of CAD software for lung nodule detection in chest computed tomography images and cerebral aneurysm detection in magnetic resonance angiography images. Twelve radiologists with different years of experience independently annotated the lesions, and the performance changes were investigated by repeating the retraining of the CAD software twice, with the addition of cases annotated by each radiologist. Additionally, we investigated the effects of retraining using integrated annotations from multiple radiologists.

RESULTS

The performance of the CAD software after retraining differed among annotating radiologists. In some cases, the performance was degraded compared to that of the initial software. Retraining using integrated annotations showed different performance trends depending on the target CAD software, notably in cerebral aneurysm detection, where the performance decreased compared to using annotations from a single radiologist.

CONCLUSIONS

Although the performance of the CAD software after retraining varied among the annotating radiologists, no direct correlation with their experience was found. The performance trends differed according to the type of CAD software used when integrated annotations from multiple radiologists were used.

Improved identification of tumors in 18F-FDG-PET examination by normalizing the standard uptake in the liver based on blood test data

PURPOSE

Standardized uptake values (SUVs) derived from 18F-fluoro-2-deoxy-D-glucose positron emission tomography/computed tomography are a crucial parameter for identifying tumors or abnormalities in an organ. Moreover, exploring ways to improve the identification of tumors or abnormalities using a statistical measurement tool is important in clinical research. Therefore, we developed a fully automatic method to create a personally normalized Z-score map of the liver SUV.

METHODS

The normalized Z-score map for each patient was created using the SUV mean and standard deviation estimated from blood-test-derived variables, such as alanine aminotransferase and aspartate aminotransferase, as well as other demographic information. This was performed using the least absolute shrinkage and selection operator (LASSO)-based estimation formula. We also used receiver operating characteristic (ROC) to analyze the results of people with and without hepatic tumors and compared them to the ROC curve of normal SUV.

RESULTS

A total of 7757 people were selected for this study. Of these, 7744 were healthy, while 13 had abnormalities. The area under the ROC curve results indicated that the anomaly detection approach (0.91) outperformed only the maximum SUV (0.89). To build the LASSO regression, sets of covariates, including sex, weight, body mass index, blood glucose level, triglyceride, total cholesterol, γ-glutamyl transpeptidase, total protein, creatinine, insulin, albumin, and cholinesterase, were used to determine the SUV mean, whereas weight was used to determine the SUV standard deviation.

CONCLUSION

The Z-score normalizes the mean and standard deviation. It is effective in ROC curve analysis and increases the clarity of the abnormality. This normalization is a key technique for effective measurement of maximum glucose consumption by tumors in the liver.

Development and evaluation of an integrated liver nodule diagnostic method by combining the liver segment division and lesion localization/classification models for enhanced focal liver lesion detection

The purpose of the study was to develop a liver nodule diagnostic method that accurately localizes and classifies focal liver lesions and identifies the specific liver segments in which they reside by integrating a liver segment division algorithm using a four-dimensional (4D) fully convolutional residual network (FC-ResNet) with a localization and classification model. We retrospectively collected data and divided 106 gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid-enhanced magnetic resonance examinations into Case-sets 1, 2, and 3. A liver segment division algorithm was developed using a 4D FC-ResNet and trained with semi-automatically created silver-standard annotations; performance was evaluated using manually created gold-standard annotations by calculating the Dice scores for each liver segment. The performance of the liver nodule diagnostic method was assessed by comparing the results with those of the original radiology reports. The mean Dice score between the output of the liver segment division model and the gold standard was 0.643 for Case-set 2 (normal liver contours) and 0.534 for Case-set 1 (deformed liver contours). Among the 64 lesions in Case-set 3, the diagnostic method localized 37 lesions, classified 33 lesions, and identified the liver segments for 30 lesions. A total of 28 lesions were true positives, matching the original radiology reports. The liver nodule diagnostic method, which integrates a liver segment division algorithm with a lesion localization and classification model, exhibits great potential for localizing and classifying focal liver lesions and identifying the liver segments in which they reside. Further improvements and validation using larger sample sizes will enhance its performance and clinical applicability.

Relationship between Thyroid CT Density, Volume, and Future TSH Elevation: A 5-Year Follow-Up Study

This study aimed to explore the relationship between thyroid-stimulating hormone (TSH) elevation and the baseline computed tomography (CT) density and volume of the thyroid. We examined 86 cases with new-onset hypothyroidism (TSH > 4.5 IU/mL) and 1071 controls from a medical check-up database over 5 years. A deep learning-based thyroid segmentation method was used to assess CT density and volume. Statistical tests and logistic regression were employed to determine differences and odds ratios. Initially, the case group showed a higher CT density (89.8 vs. 81.7 Hounsfield units (HUs)) and smaller volume (13.0 vs. 15.3 mL) than those in the control group. For every +10 HU in CT density and -3 mL in volume, the odds of developing hypothyroidism increased by 1.40 and 1.35, respectively. Over the course of the study, the case group showed a notable CT density reduction (median: -8.9 HU), whereas the control group had a minor decrease (-2.9 HU). Thyroid volume remained relatively stable for both groups. Higher CT density and smaller thyroid volume at baseline are correlated with future TSH elevation. Over time, there was a substantial and minor decrease in CT density in the case and control groups, respectively. Thyroid volumes remained consistent in both cohorts.

[A Case of Sigmoid Colon Cancer with Liver Metastasis and Gastric Infiltration for Which Ileus Surgery Was the Impetus for Surgery]

The case is a woman in her 60s. Sigmoid colon cancer surgery, liver metastasis surgery, and adjuvant chemotherapy were performed at another hospital 2 years ago. Later, she developed a metastasis in her liver and was recommended surgery, but she refused treatment and was transferred. Her liver metastasis had invaded the stomach and formed a giant gastric ulcer. This time she had an adhesive ileus and underwent laparoscopic surgery at our hospital. At that time, we observed the state of liver metastasis and gastric infiltration by laparoscopy, so we thought that palliative surgery was possible and recommended it. Although she initially refused treatment, the relative ease with which her ileus surgery was performed encouraged her to undergo palliative surgery. Laparoscopic-assisted gastrectomy and partial hepatectomy were performed, and she was discharged on hospital day 13 after surgery. She subsequently developed liver metastases and died 8 months after palliative surgery, although she was able to eat and maintain her ADL until the end of life. By staying close to the patient, we were able to lead the patient from refusal of surgery to palliative surgery, and we felt that we were able to make the patient reach a favorable end.

Laminar-type gratings overcoated with carbon-based materials to enhance analytical sensitivity of flat-field emission spectrograph in the VUV region

Laminar-type spherical diffraction gratings overcoated with carbon-based materials were designed, fabricated, and evaluated for the purpose of enhancing the analytical sensitivity of the flat-field spectrograph in a vacuum ultraviolet region of 35-110 eV. As the design benchmark for numerical calculations, diffraction efficiency (DE) and spectral flux, which are defined by the product of the DE and numerical aperture and correlate with the analytical sensitivity of the spectrograph, were used. To simplify the feasibility study on the overcoating effects, we assumed a laminar-type grating having a grating constant of 1/1000 mm and coated with a Au layer of 30.0 nm thickness and an incidence angle of 84.0°. The optimized groove depth and duty ratio were 30.0 nm and 0.3, respectively. In addition, the optimum thicknesses of the overcoating layer were 44, 46, 24, and 30 nm for B4C, C, diamond-like-carbon, and SiC, respectively. Based on these results, we have fabricated a varied-line-spacing holographic grating overcoated with B4C with a thickness of 47 nm. For the experimental evaluation, we used the light source of Mg-L and Al-L emissions excited by the electron beam generated from an electron microscope, an objective flat-field spectrograph, and a CCD imaging detector. The experimental results showed that the spectrograph employing a new grating overcoated with the B4C layer indicated almost the same spectral resolution and 2.9-4.2 times higher analytical sensitivity compared with those obtained with a previously designed Au-coated grating having a grating constant of 1/1200 mm and used at an incidence of 86.0°.

Synthesis, structures, and solid-state photoresponsive color change behavior of boronium complexes bearing a pyridine-imine, diimine, or pyridine-ketone bidentate ligand

9-Borabicyclo[3.3.1]nonane-based boronium triflates bearing a N-substituted 2-pyridylmethanimine, N,N'-dialkylethane-1,2-diimine, or 2-arylcarbonylpyridine ligand were synthesized. Their tetracoordinate boron structures were determined using 11B NMR spectra and X-ray crystallography. The pyridine-imine complexes exhibited solid-state photoresponsive color changes upon UV irradiation, which indicated that boronium complexes without a bipyridine moiety also have photoresponsive capabilities. Combination of TD-DFT calculations and measurements of UV-vis absorption and fluorescence properties, diffuse reflectance spectra, and ESR spectra provided suggestions on the determining factor of the photoresponsive color change capabilities and structures of the photoproducts.

Retrospective study to estimate the prevalence and describe the clinicopathological characteristics, treatments received, and outcomes of HER2-low breast cancer

BACKGROUND

Approximately 80% of all breast cancers (BCs) are currently categorized as human epidermal growth factor receptor 2 (HER2)-negative [immunohistochemistry (IHC) 0, 1+, or 2+/in situ hybridization (ISH) negative]; approximately 60% of BCs traditionally categorized as HER2-negative express low levels of HER2. HER2-low (IHC 1+ or IHC 2+/ISH-) status became clinically actionable with approval of trastuzumab deruxtecan to treat unresectable/metastatic HER2-low BC. Greater understanding of patients with HER2-low disease is urgently needed.

PATIENTS AND METHODS

This global, multicenter, retrospective study (NCT04807595) included tissue samples from patients with confirmed HER2-negative unresectable/metastatic BC [any hormone receptor (HR) status] diagnosed from 2014 to 2017. Pathologists rescored HER2 IHC-stained slides as HER2-low (IHC 1+ or IHC 2+/ISH-) or HER2 IHC 0 after training on low-end expression scoring using Ventana 4B5 and other assays at local laboratories (13 sites; 10 countries) blinded to historical scores. HER2-low prevalence and concordance between historical scores and rescores were assessed. Demographics, clinicopathological characteristics, treatments, and outcomes were examined.

RESULTS

In rescored samples from 789 patients with HER2-negative unresectable/metastatic BC, the overall HER2-low prevalence was 67.2% (HR positive, 71.1%; HR negative, 52.8%). Concordance was moderate between historical and rescored HER2 statuses (81.3%; κ = 0.583); positive agreement was numerically higher for HER2-low (87.5%) than HER2 IHC 0 (69.9%). More than 30% of historical IHC 0 cases were rescored as HER2-low overall (all assays) and using Ventana 4B5. There were no notable differences between HER2-low and HER2 IHC 0 in patient characteristics, treatments received, or clinical outcomes.

CONCLUSIONS

Approximately two-thirds of patients with historically HER2-negative unresectable/metastatic BC may benefit from HER2-low-directed treatments. Our data suggest that HER2 reassessment in patients with historical IHC 0 scores may be considered to help optimize selection of patients for treatment. Further, accurate identification of patients with HER2-low BC may be achieved with standardized pathologist training.

Characterization of Brain Volume Changes in Aging Individuals With Normal Cognition Using Serial Magnetic Resonance Imaging

Importance

Characterizing longitudinal patterns of regional brain volume changes in a population with normal cognition at the individual level could improve understanding of the brain aging process and may aid in the prevention of age-related neurodegenerative diseases.

Objective

To investigate age-related trajectories of the volumes and volume change rates of brain structures in participants without dementia.

Design, Setting, and Participants

This cohort study was conducted from November 1, 2006, to April 30, 2021, at a single academic health-checkup center among 653 individuals who participated in a health screening program with more than 10 years of serial visits.

Exposure

Serial magnetic resonance imaging, Mini-Mental State Examination, health checkup.

Main Outcomes and Measures

Volumes and volume change rates across brain tissue types and regions.

Results

The study sample included 653 healthy control individuals (mean [SD] age at baseline, 55.1 [9.3] years; median age, 55 years [IQR, 47-62 years]; 447 men [69%]), who were followed up annually for up to 15 years (mean [SD], 11.5 [1.8] years; mean [SD] number of scans, 12.1 [1.9]; total visits, 7915). Each brain structure showed characteristic age-dependent volume and atrophy change rates. In particular, the cortical gray matter showed a consistent pattern of volume loss in each brain lobe with aging. The white matter showed an age-related decrease in volume and an accelerated atrophy rate (regression coefficient, -0.016 [95% CI, -0.012 to -0.011]; P < .001). An accelerated age-related volume increase in the cerebrospinal fluid-filled spaces, particularly in the inferior lateral ventricle and the Sylvian fissure, was also observed (ventricle regression coefficient, 0.042 [95% CI, 0.037-0.047]; P < .001; sulcus regression coefficient, 0.021 [95% CI, 0.018-0.023]; P < .001). The temporal lobe atrophy rate accelerated from approximately 70 years of age, preceded by acceleration of atrophy in the hippocampus and amygdala.

Conclusions and Relevance

In this cohort study of adults without dementia, age-dependent brain structure volumes and volume change rates in various brain structures were characterized using serial magnetic resonance imaging scans. These findings clarified the normal distributions in the aging brain, which are essential for understanding the process of age-related neurodegenerative diseases.

Significance of FDG-PET standardized uptake values in predicting thyroid disease

OBJECTIVE

This study aimed to determine a standardized cut-off value for abnormal 18F-fluorodeoxyglucose (FDG) accumulation in the thyroid gland.

METHODS

Herein, 7013 FDG-PET/CT scans were included. An automatic thyroid segmentation method using two U-nets (2D- and 3D-U-net) was constructed; mean FDG standardized uptake value (SUV), CT value, and volume of the thyroid gland were obtained from each participant. The values were categorized by thyroid function into three groups based on serum thyroid-stimulating hormone levels. Thyroid function and mean SUV with increments of 1 were analyzed, and risk for thyroid dysfunction was calculated. Thyroid dysfunction detection ability was examined using a machine learning method (LightGBM, Microsoft) with age, sex, height, weight, CT value, volume, and mean SUV as explanatory variables.

RESULTS

Mean SUV was significantly higher in females with hypothyroidism. Almost 98.9% of participants in the normal group had mean SUV < 2 and 93.8% participants with mean SUV < 2 had normal thyroid function. The hypothyroidism group had more cases with mean SUV ≥ 2. The relative risk of having abnormal thyroid function was 4.6 with mean SUV ≥ 2. The sensitivity and specificity for detecting thyroid dysfunction using LightGBM (Microsoft) were 14.5 and 99%, respectively.

CONCLUSIONS

Mean SUV ≥ 2 was strongly associated with abnormal thyroid function in this large cohort, indicating that mean SUV with FDG-PET/CT can be used as a criterion for thyroid evaluation. Preliminarily, this study shows the potential utility of detecting thyroid dysfunction based on imaging findings.

Axillary Lymphadenopathy after Pfizer-BioNTech and Moderna COVID-19 Vaccination: MRI Evaluation

Background COVID-19 vaccination-related axillary lymphadenopathy has become an important problem in cancer imaging. Data are needed to update or support imaging guidelines for conducting appropriate follow-up. Purpose To investigate the prevalence, predisposing factors, and MRI characteristics of COVID-19 vaccination-related axillary lymphadenopathy. Materials and Methods Prospectively collected prevaccination and postvaccination chest MRI scans were secondarily analyzed. Participants who underwent two doses of either the Pfizer-BioNTech or Moderna COVID-19 vaccine and chest MRI from June to October 2021 were included. Enlarged axillary lymph nodes were identified on postvaccination MRI scans compared with prevaccination scans. The lymph node diameter, signal intensity with T2-weighted imaging, and apparent diffusion coefficient (ADC) of the largest enlarged lymph nodes were measured. These values were compared between prevaccination and postvaccination MRI by using the Wilcoxon signed-rank test. Results Overall, 433 participants (mean age, 65 years ± 11 [SD]; 300 men and 133 women) were included. The prevalence of axillary lymphadenopathy in participants 1-14 days after vaccination was 65% (30 of 46). Participants with lymphadenopathy were younger than those without lymphadenopathy (P < .001). Female sex and the Moderna vaccine were predisposing factors (P = .005 and P = .003, respectively). Five or more enlarged lymph nodes were noted in 2% (eight of 433) of participants. Enlarged lymph nodes greater than or equal to 10 mm in the short axis were noted in 1% (four of 433) of participants. The median signal intensity relative to the muscle on T2-weighted images was 4.0; enlarged lymph nodes demonstrated a higher signal intensity (P = .002). The median ADC of enlarged lymph nodes after vaccination in 90 participants was 1.1 × 10^-3 mm²/sec (range, 0.6-2.0 × 10^-3 mm²/sec), thus ADC values remained normal. Conclusion Axillary lymphadenopathy after the second dose of the Pfizer-BioNTech or Moderna COVID-19 vaccines was frequent within 2 weeks after vaccination, was typically less than 10 mm in size, and had a normal apparent diffusion coefficient. © RSNA, 2022.

On the Simulation of Ultra-Sparse-View and Ultra-Low-Dose Computed Tomography with Maximum a Posteriori Reconstruction Using a Progressive Flow-Based Deep Generative Model

Ultra-sparse-view computed tomography (CT) algorithms can reduce radiation exposure for patients, but these algorithms lack an explicit cycle consistency loss minimization and an explicit log-likelihood maximization in testing. Here, we propose X2CT-FLOW for the maximum a posteriori (MAP) reconstruction of a three-dimensional (3D) chest CT image from a single or a few two-dimensional (2D) projection images using a progressive flow-based deep generative model, especially for ultra-low-dose protocols. The MAP reconstruction can simultaneously optimize the cycle consistency loss and the log-likelihood. We applied X2CT-FLOW for the reconstruction of 3D chest CT images from biplanar projection images without noise contamination (assuming a standard-dose protocol) and with strong noise contamination (assuming an ultra-low-dose protocol). We simulated an ultra-low-dose protocol. With the standard-dose protocol, our images reconstructed from 2D projected images and 3D ground-truth CT images showed good agreement in terms of structural similarity (SSIM, 0.7675 on average), peak signal-to-noise ratio (PSNR, 25.89 dB on average), mean absolute error (MAE, 0.02364 on average), and normalized root mean square error (NRMSE, 0.05731 on average). Moreover, with the ultra-low-dose protocol, our images reconstructed from 2D projected images and the 3D ground-truth CT images also showed good agreement in terms of SSIM (0.7008 on average), PSNR (23.58 dB on average), MAE (0.02991 on average), and NRMSE (0.07349 on average).

Clinical Comparable Corpus Describing the Same Subjects with Different Expressions

Medical artificial intelligence (AI) systems need to learn to recognize synonyms or paraphrases describing the same anatomy, disease, treatment, etc. to better understand real-world clinical documents. Existing linguistic resources focus on variants at the word or sentence level. To handle linguistic variations on a broader scale, we proposed the Medical Text Radiology Report section Japanese version (MedTxt-RR-JA), the first clinical comparable corpus. MedTxt-RR-JA was built by recruiting nine radiologists to diagnose the same 15 lung cancer cases in Radiopaedia, an open-access radiological repository. The 135 radiology reports in MedTxt-RR-JA were shown to contain word-, sentence- and document-level variations maintaining similarity of contents. MedTxt-RR-JA is also the first publicly available Japanese radiology report corpus that would help to overcome poor data availability for Japanese medical AI systems. Moreover, our methodology can be applied widely to building clinical corpora without privacy concerns.

Anomaly detection in chest 18F-FDG PET/CT by Bayesian deep learning

Purpose

To develop an anomaly detection system in PET/CT with the tracer ¹⁸F-fluorodeoxyglucose (FDG) that requires only normal PET/CT images for training and can detect abnormal FDG uptake at any location in the chest region.

Materials and methods

We trained our model based on a Bayesian deep learning framework using 1878 PET/CT scans with no abnormal findings. Our model learns the distribution of standard uptake values in these normal training images and detects out-of-normal uptake regions. We evaluated this model using 34 scans showing focal abnormal FDG uptake in the chest region. This evaluation dataset includes 28 pulmonary and 17 extrapulmonary abnormal FDG uptake foci. We performed per-voxel and per-slice receiver operating characteristic (ROC) analyses and per-lesion free-response receiver operating characteristic analysis.

Results

Our model showed an area under the ROC curve of 0.992 on discriminating abnormal voxels and 0.852 on abnormal slices. Our model detected 41 of 45 (91.1%) of the abnormal FDG uptake foci with 12.8 false positives per scan (FPs/scan), which include 26 of 28 pulmonary and 15 of 17 extrapulmonary abnormalities. The sensitivity at 3.0 FPs/scan was 82.2% (37/45).

Conclusion

Our model trained only with normal PET/CT images successfully detected both pulmonary and extrapulmonary abnormal FDG uptake in the chest region.

Preliminary study of generalized semiautomatic segmentation for 3D voxel labeling of lesions based on deep learning

PURPOSE

The three-dimensional (3D) voxel labeling of lesions requires significant radiologists' effort in the development of computer-aided detection software. To reduce the time required for the 3D voxel labeling, we aimed to develop a generalized semiautomatic segmentation method based on deep learning via a data augmentation-based domain generalization framework. In this study, we investigated whether a generalized semiautomatic segmentation model trained using two types of lesion can segment previously unseen types of lesion.

METHODS

We targeted lung nodules in chest CT images, liver lesions in hepatobiliary-phase images of Gd-EOB-DTPA-enhanced MR imaging, and brain metastases in contrast-enhanced MR images. For each lesion, the 32 × 32 × 32 isotropic volume of interest (VOI) around the center of gravity of the lesion was extracted. The VOI was input into a 3D U-Net model to define the label of the lesion. For each type of target lesion, we compared five types of data augmentation and two types of input data.

RESULTS

For all considered target lesions, the highest dice coefficients among the training patterns were obtained when using a combination of the existing data augmentation-based domain generalization framework and random monochrome inversion and when using the resized VOI as the input image. The dice coefficients were 0.639 ± 0.124 for the lung nodules, 0.660 ± 0.137 for the liver lesions, and 0.727 ± 0.115 for the brain metastases.

CONCLUSIONS

Our generalized semiautomatic segmentation model could label unseen three types of lesion with different contrasts from the surroundings. In addition, the resized VOI as the input image enables the adaptation to the various sizes of lesions even when the size distribution differed between the training set and the test set.

Versatile anomaly detection method for medical images with semi-supervised flow-based generative models

PURPOSE

Radiologists interpret many medical images and clinical practice demands timely interpretation, resulting in a heavy workload. To reduce the workload, here we formulate and validate a method that can handle different types of medical image and can detect virtually all types of lesion in a medical image. For the first time, we show that two flow-based deep generative (FDG) models can predict the logarithm posterior probability in a semi-supervised approach.

METHODS

We adopt two FDG models in conjunction with Bayes' theorem to predict the logarithm posterior probability that a medical image is normal. We trained one of the FDG models with normal images and the other FDG model with normal and non-normal images.

RESULTS

We validated the method using two types of medical image: chest X-ray images (CXRs) and brain computed tomography images (BCTs). The area under the receiver operating characteristic curve for pneumonia-like opacities in CXRs was 0.839 on average, and for infarction in BCTs was 0.904.

CONCLUSION

We formulated a method of predicting the logarithm posterior probability using two FDG models. We validated that the method can detect abnormal findings in CXRs and BCTs with both an acceptable performance for testing and a comparatively light workload for training.

Performance changes due to differences in training data for cerebral aneurysm detection in head MR angiography images

PURPOSE

The performance of computer-aided detection (CAD) software depends on the quality and quantity of the dataset used for machine learning. If the data characteristics in development and practical use are different, the performance of CAD software degrades. In this study, we investigated changes in detection performance due to differences in training data for cerebral aneurysm detection software in head magnetic resonance angiography images.

MATERIALS AND METHODS

We utilized three types of CAD software for cerebral aneurysm detection in MRA images, which were based on 3D local intensity structure analysis, graph-based features, and convolutional neural network. For each type of CAD software, we compared three types of training pattern, which were two types of training using single-site data and one type of training using multisite data. We also carried out internal and external evaluations.

RESULTS

In training using single-site data, the performance of CAD software largely and unpredictably fluctuated when the training dataset was changed. Training using multisite data did not show the lowest performance among the three training patterns for any CAD software and dataset.

CONCLUSION

The training of cerebral aneurysm detection software using data collected from multiple sites is desirable to ensure the stable performance of the software.

Multichannel three-dimensional fully convolutional residual network-based focal liver lesion detection and classification in Gd-EOB-DTPA-enhanced MRI

PURPOSE

Gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid (Gd-EOB-DTPA)-enhanced magnetic resonance imaging (MRI) has high diagnostic accuracy in the detection of liver lesions. There is a demand for computer-aided detection/diagnosis software for Gd-EOB-DTPA-enhanced MRI. We propose a deep learning-based method using one three-dimensional fully convolutional residual network (3D FC-ResNet) for liver segmentation and another 3D FC-ResNet for simultaneous detection and classification of a focal liver lesion in Gd-EOB-DTPA-enhanced MRI.

METHODS

We prepared a five-phase (unenhanced, arterial, portal venous, equilibrium, and hepatobiliary phases) series as the input image sets and labeled focal liver lesion (hepatocellular carcinoma, metastasis, hemangiomas, cysts, and scars) images as the output image sets. We used 100 cases to train our model, 42 cases to determine the hyperparameters of our model, and 42 cases to evaluate our model. We evaluated our model by free-response receiver operating characteristic curve analysis and using a confusion matrix.

RESULTS

Our model simultaneously detected and classified focal liver lesions. In the test cases, the detection accuracy for whole focal liver lesions had a true-positive ratio of 0.6 at an average of 25 false positives per case. The classification accuracy was 0.790.

CONCLUSION

We proposed the simultaneous detection and classification of a focal liver lesion in Gd-EOB-DTPA-enhanced MRI using multichannel 3D FC-ResNet. Our results indicated simultaneous detection and classification are possible using a single network. It is necessary to further improve detection sensitivity to help radiologists.

Unsupervised Deep Anomaly Detection in Chest Radiographs

The purposes of this study are to propose an unsupervised anomaly detection method based on a deep neural network (DNN) model, which requires only normal images for training, and to evaluate its performance with a large chest radiograph dataset. We used the auto-encoding generative adversarial network (α-GAN) framework, which is a combination of a GAN and a variational autoencoder, as a DNN model. A total of 29,684 frontal chest radiographs from the Radiological Society of North America Pneumonia Detection Challenge dataset were used for this study (16,880 male and 12,804 female patients; average age, 47.0 years). All these images were labeled as "Normal," "No Opacity/Not Normal," or "Opacity" by board-certified radiologists. About 70% (6,853/9,790) of the Normal images were randomly sampled as the training dataset, and the rest were randomly split into the validation and test datasets in a ratio of 1:2 (7,610 and 15,221). Our anomaly detection system could correctly visualize various lesions including a lung mass, cardiomegaly, pleural effusion, bilateral hilar lymphadenopathy, and even dextrocardia. Our system detected the abnormal images with an area under the receiver operating characteristic curve (AUROC) of 0.752. The AUROCs for the abnormal labels Opacity and No Opacity/Not Normal were 0.838 and 0.704, respectively. Our DNN-based unsupervised anomaly detection method could successfully detect various diseases or anomalies in chest radiographs by training with only the normal images.

Computer-aided detection of cerebral aneurysms with magnetic resonance angiography: usefulness of volume rendering to display lesion candidates

PURPOSE

The clinical usefulness of computer-aided detection of cerebral aneurysms has been investigated using different methods to present lesion candidates, but suboptimal methods may have limited its usefulness. We compared three presentation methods to determine which can benefit radiologists the most by enabling them to detect more aneurysms.

MATERIALS AND METHODS

We conducted a multireader multicase observer performance study involving six radiologists and using 470 lesion candidates output by a computer-aided detection program, and compared the following three different presentation methods using the receiver operating characteristic analysis: (1) a lesion candidate is encircled on axial slices, (2) a lesion candidate is overlaid on a volume-rendered image, and (3) combination of (1) and (2). The response time was also compared.

RESULTS

As compared with axial slices, radiologists showed significantly better detection performance when presented with volume-rendered images. There was no significant difference in response time between the two methods. The combined method was associated with a significantly longer response time, but had no added merit in terms of diagnostic accuracy.

CONCLUSION

Even with the aid of computer-aided detection, radiologists overlook many aneurysms if the presentation method is not optimal. Overlaying colored lesion candidates on volume-rendered images can help them detect more aneurysms.

Anatomical study of cervical intervertebral foramen in patients with cervical spondylotic radiculopathy

PURPOSE

To determine the underlying anatomical characteristics in patients with cervical spondylotic radiculopathy (CSR) by comparing those of surgically treated CSR patients with those of healthy subjects.

METHODS

Computed tomography (CT) scans of the cervical spine in 42 patients who underwent decompression surgery for CSR were investigated. As a control group, 42 age- and sex-matched healthy subjects were randomly selected from the 1272 subjects who underwent CT examinations of the entire spine as their routine medical check-up. Image measurements included C2-7 sagittal Cobb angle, spinal canal diameters, and angles of the nerve root groove at each level from C3 to C7, and the size of the intervertebral foramen and the size of osteophytes at each level from C3/4 to C7/T1. As for the frequency of osteophytes at the surgical level, we compared the operated and nonoperated intervertebral foramina among the CSR patients, and all other parameters were compared with the corresponding segments in the control group.

RESULTS

Forty-eight intervertebral segments were surgically treated in the CSR group. There was a higher incidence of osteophytes in the operated foramen (70.8%) than in the nonoperated foramen (28.2%, p < 0.01) in the patients with CSR. The anteroposterior diameter (width) of the foramen was significantly smaller at all levels in the CSR patients, whereas the height of the foramen did not significantly differ between the two groups.

CONCLUSION

It can be speculated that the width of the intervertebral foramen (developmental factor) and the formation of osteophytes (spondylotic factor) were related to the onset of the CSR.

Effects of π-conjugation on the solid-state photoresponsive coloring behavior of bipyridine-boronium complexes

Solid-state photoinduced coloring of boronium complexes consisting of 9-borabicyclononane and 2,2'-bipyridine with π-conjugated substituents at the 4,4'- or 5,5'-positions was investigated. The substitution position affected the highest occupied molecular orbital distribution and determined the coloring capability. The 4,4'-substituted complexes exhibited coloration upon irradiation, whereas most of the 5,5'-substituted complexes did not.

Association between insulin resistance and left ventricular diastolic dysfunction in non-diabetic general population

Background

Insulin resistance due to visceral fat accumulation plays a central role in the development of diabetic mellitus and subsequent cardiovascular disease. Abdominal obesity and diabetic mellitus are associated with the progression of left ventricular diastolic dysfunction which is the major structural abnormality in patients with heart failure. However, whether insulin resistance influences left ventricular diastolic dysfunction in non-diabetic subjects is unclear.

Purpose

In this study, we aimed to clarify the association between insulin resistance and left ventricular diastolic dysfunction in non-diabetic general population.

Methods

We examined 2,572 non-diabetic subjects with preserved left ventricular systolic function (ejection fraction ≥50%) and without a past history of diabetic mellitus, heart failure, coronary artery disease, atrial fibrillation, stroke, moderate to severe valvular disease, who underwent medical check-ups at the University of Tokyo Hospital from January 2009 to December 2018. Diabetes mellitus was defined as fasting glucose level ≥126 mg/dL or a subject's use of oral antidiabetic medications or insulin. We calculated the Homeostasis Model Assessment-Insulin Resistance (HOMA-IR) score [HOMA-IR = fasting glucose (mg/dL) × insulin (μU/mL) / 405], and defined insulin resistance as HOMA-IR ≥2.5. Left ventricular diastolic function was assessed by echocardiography, using tissue doppler analysis (E/E' ratio in septal and lateral). Visceral adiposity was assessed as visceral fat volume measured by computed tomography scanner.

Results

Mean age was 53.3±9.8 years, and 1,746 subjects (67.9%) were male. Overall, HOMA-IR was 1.1 on average, and HOMA-IR ≥2.5 was observed in 5.0% of study population. BMI and the prevalence of obesity were higher in subjects with HOMA-IR ≥2.5 than in those without. There was no significant difference in left ventricular ejection fraction between HOMA-IR ≥2.5 and ≤2.5. E/E' ratio in septal and lateral were higher in subjects with HOMA-IR ≥2.5 than in those without. Multivariable logistic regression analysis demonstrated that obesity and waist circumference, visceral fat volume were independently associated with HOMA-IR ≥2.5, however, age, hypertension, hypercholesterolemia were not. Further, multiple regression analysis including insulin resistance, age, male gender, obesity, hypertension, hypercholesterolemia, cigarette smoking and visceral fat volume, showed that insulin resistance was an independent determinant of increasing E/E' in both septal and lateral, whereas visceral fat volume was not (Figure 1).

Conclusion

The results of the present study suggest that insulin resistance due to visceral fat accumulation might be associated with the development of left ventricular diastolic dysfunction in non-diabetic general population without overt cardiovascular disease.

Figure 1

Funding Acknowledgement

Type of funding source: None

Risk factors for progression of ossification of the posterior longitudinal ligament in asymptomatic subjects

OBJECTIVE

The incidence and risk factors for the progression of ossification of the posterior longitudinal ligament (OPLL) have been previously reported in surgically and nonsurgically treated symptomatic patients. However, the correlates of OPLL progression in asymptomatic subjects with OPLL are not well characterized. This study aimed to clarify the incidence and risk factors for OPLL progression in asymptomatic subjects based on whole-body CT.

METHODS

The authors retrospectively reviewed 2585 healthy subjects who underwent whole-body CT at a single health center from September 2007 to December 2011. This study included asymptomatic subjects with OPLL who underwent CT scans twice with an interval of at least 5 years. Progression of OPLL was assessed based on initial and final CT scan. Subjects were divided into two groups: nonprogression (OPLL-NP) and progression (OPLL-P) groups. Clinical characteristics, bone mineral density status, OPLL types, and OPLL involvement of multiple vertebral levels between the two groups were compared. Risk factors for progression of OPLL were identified by logistic regression analysis after propensity score adjustment.

RESULTS

Of the 109 subjects with OPLL (91 men and 18 women), 20 (18.3%) exhibited OPLL progression (OPLL-P group). Subjects in the OPLL-P group were significantly younger (p = 0.031), had higher prevalence of multilevel OPLL involvement (p = 0.041) and continuous type of OPLL (p = 0.015), and had higher uric acid (UA) levels (p = 0.004) at the time of initial CT examination compared to the OPLL-NP group. Younger age (adjusted odds ratio [aOR] 0.95, 95% CI 0.90-0.99), OPLL involvement of multiple vertebral levels (aOR 2.88, 95% CI 1.06-7.83), continuous type of OPLL (aOR 4.21, 95% CI 1.35-13.10), and higher UA levels (aOR 2.09, 95% CI 1.24-3.53) were significant risk factors for OPLL progression.

CONCLUSIONS

Younger age, OPLL involvement of multiple vertebral levels, continuous type of OPLL, and higher UA levels are significant risk factors for OPLL progression in asymptomatic subjects.

Bone strength of the proximal femur in healthy subjects with ossification of the posterior longitudinal ligament

We compared the bone strength measured via quantitative computed tomography-based finite element method (QCT/FEM) between healthy adults with and without ossification of the posterior longitudinal ligament (OPLL). No statistically significant difference was observed in the bone strength between healthy adults with and without OPLL. Hyperostosis of the posterior longitudinal ligament in OPLL may not be associated with the systemic bone strength.

INTRODUCTION

Although patients with OPLL have been reportedly associated with increased level of bone mineral density (BMD) using dual-energy X-ray absorptiometry (DXA), little is known about the bone strength in OPLL subjects. The aim of this study is to investigate the bone strength measured via QCT/FEM in healthy subjects with OPLL using the medical check-up data, including whole-body CT scans.

METHODS

We examined 796 participants (529 men and 267 women) who underwent CT scans in a single health center between January 2008 and May 2009. We identified OPLL in whole spine and divided the subjects into two groups: non-OPLL and OPLL groups. We calculated the predicted bone strength (PBS) of the proximal femur using QCT/FEM and examined the bone mineral status of the calcaneus using quantitative ultrasound (QUS). We compared the PBS and the QUS parameters between the non-OPLL and OPLL groups.

RESULTS

Seventy-four subjects (9.3%; 57 men and 17 women) were diagnosed with OPLL in the whole spine. The OPLL group was significantly older than the non-OPLL group. No statistically significant difference was observed in the PBS and the QUS parameters between the non-OPLL and OPLL groups in both sexes. Furthermore, no statistically significant difference was noted in the PBS and the QUS parameters between two groups in age- and gender-matched analysis.

CONCLUSIONS

Our results suggest that hyperostosis of the posterior longitudinal ligament in OPLL may not be associated with bone strength and bone mineral status at the extremities.

Novel platform for development, training, and validation of computer-assisted detection/diagnosis software

PURPOSE

To build a novel, open-source, purely web-based platform system to address problems in the development and clinical use of computer-assisted detection/diagnosis (CAD) software. The new platform system will replace the existing system for the development and validation of CAD software, Clinical Infrastructure for Radiologic Computation of United Solutions (CIRCUS).

METHODS

In our new system, the two top-level applications visible to users are the web-based image database (CIRCUS DB; database) and the Docker plug-in-based CAD execution platform (CIRCUS CS; clinical server). These applications are built on top of a shared application programming interface server, a three-dimensional image viewer component, and an image repository.

RESULTS

We successfully installed our new system into a Linux server at two clinical sites. A total of 1954 cases were registered in CIRCUS DB. We have been utilizing CIRCUS CS with four Docker-based CAD plug-ins.

CONCLUSIONS

We have successfully built a new version of the CIRCUS system. Our platform was successfully implemented at two clinical sites, and we plan to publish it as an open-source software project.

Pilot study of eruption forecasting with muography using convolutional neural network

Muography is a novel method of visualizing the internal structures of active volcanoes by using high-energy near-horizontally arriving cosmic muons. The purpose of this study is to show the feasibility of muography to forecast the eruption event with the aid of the convolutional neural network (CNN). In this study, seven daily consecutive muographic images were fed into the CNN to compute the probability of eruptions on the eighth day, and our CNN model was trained by hyperparameter tuning with the Bayesian optimization algorithm. By using the data acquired in Sakurajima volcano, Japan, as an example, the forecasting performance achieved a value of 0.726 for the area under the receiver operating characteristic curve, showing the reasonable correlation between the muographic images and eruption events. Our result suggests that muography has the potential for eruption forecasting of volcanoes.

Prediction of the pathological fracture risk during stance and fall-loading configurations for metastases in the proximal femur, using a computed tomography-based finite element method

BACKGROUND

It is important to assess the fracture risk associated with metastasis in the proximal femur. The study aimed to clarify the effect of tumor location on the risk of pathological fracture of the proximal femur and investigate the fracture risk not only in the stance-loading configuration (SC), but also in the fall-loading configuration (FC) using a computed tomography (CT)-based finite element (FE) method based on a simulated metastatic model.

METHODS

The axial CT scans of the proximal femora of non-osteoporotic healthy men (n = 4; age range, 42-48 years) and osteoporotic post-menopausal women (n = 4; age range, 69-78 years) were obtained with a calibration phantom, from which the three-dimensional FE models were constructed. A single 15-mm-diameter spherical void simulating a tumor was created at various locations from the neck to subtrochanteric level. Nonlinear FE analyses were performed.

RESULTS

The mean predicted fracture loads without spherical voids in the SC were 7700 N in men and 4370 N in women. With the void at the medial femoral neck and in the region anteromedial to lesser trochanter, the mean predicted fracture load significantly reduced to 51.3% and 59.4% in men and 34.1% and 64.5% in women, respectively. The mean predicted fracture loads without a spherical void in the FC were 2500 N in men and 1862 N in women. With the void at the medial and posterior femoral neck, the predicted fracture load was significantly reduced to 65.7% and 79.7% in men and 48.3% and 65.4% in women, respectively.

CONCLUSIONS

These results showed that the risk of pathologic fracture was quite high in both the SC and FC when the lytic lesion existed along the principal compressive trabecular trajectory or posterior neck. Prophylactic intervention should be considered for metastases at these locations.

P3541Change in left ventricular ejection fraction and outcome in heart failure patients with mid-range ejection fraction: from the HIJ-HF prospective study

Introduction

Heart failure (HF) is categorized on the basis of the eft ventricular ejection fraction (LVEF). The European Society of Cardiology has proposed mid-range EF (HFmrEF) as a new category of HF that includes patients with an LVEF of 40–49%. However, the clinical characteristics, change in LVEF following treatment, and outcome of patients with HFmrEF remain clear.

Methods

We conducted a prospective observational study of Japanese hospitalized HF patients between 2015 and 2018 at a single-center (HIJ-HF III). HFmrEF was defined as 40–49% of LVEF on echocardiography at admission. We followed these patients and performed echocardiography to assess LVEF per year after hospital discharge. Clinical outcome was death from any cause.

Methods and results

We studied 138 patients with HFmrEF (median age 71 years, 69% male). They had 32% of ischemic heart disease, 9% of New York Heart Association functional class III or IV at discharge. During median follow-up of 20 [13–28] months, we assessed change in LVEF for 110 patients with HFmrEF. One year after hospital discharge, 49 patients (44%) improved LVEF (≥50%) and 21 patients (19%) reduced LVEF (<40%). HFmrEF patients who reduced LVEF (<40%) were significantly higher mortality rate than those who improved LVEF (≥50%) (14% vs. 2%, p<0.05) (Figure).

Conclusions

This study demonstrated that 44% of HFmrEF patients improved LVEF following treatment but 19% patients reduced LVEF. Reduced LVEF was associated with poor prognosis.

4332Benefit of sinus rhythm restoration in acute decompensated heart failure patients with atrial tachyarrhythmia treated with landiolol

Background

Atrial tachyarrhythmias (ATA), such as atrial fibrillation / atrial tachycardia are frequently observed in patients with acute decompensated heart failure (ADHF). Because ATA leads to clinical deterioration and worsen HF, the conversion and prevention of ATA is important of ADHF with ATA. Landiolol, an ultrashort-acting intravenous beta-1 blocker, was developed and has been used for the treatment of ATA.

Purpose

We evaluated the acute effect of landiolol treatment on heart rate or blood pressure (BP), also the rates and benefits of sinus rhythm (SR) restoration among AHF patients with ATA treated with landiolol.

Methods

We studied 67 consecutive HF patients with ATA (age: 67±12 years, 36 male) treated with landiolol from 2015 to December 2017 at our University Hospital. They were compared with 50 paired subjects, matched for gender, age and baseline BP who developed HF with ATA from HIJ-HF 2 study (consisted of HF patients hospitalized between 2013 and 2014).

Results

At the start of landiolol treatment, mean left ventricular ejection fraction (LVEF) was 41±14%. The median maintenance dose of landiolol was 3.0 (1.0–12.0) μ/kg/min and the median treatment duration of landiolol was 5 (1–24) days. After 6 hours from administration of landiolol, mean HR decreased significantly from 140±18 to 100±21 bpm (p<0.05), whereas BP was not difference during landiolol treatment. Sinus rhythm was restored spontaneously in 15 (22%), and by electrical or pharmacological cardioversion in 5 (7%) during a treatment with intravenous landiolol. Furthermore, sinus rhythm was restored in 22 patients using additional rhythm control treatment, such as amiodarone or catheter ablation after intravenous landiolol treatement. Eight patients experienced in-hospital death. Forty-one (69%) of 59 patients discharged alive were in SR. During the follow-up period of 16±12 months, 4 patients died and 12 patients experienced rehospitalization due to worsening HF after hospital discharge. There was a significant higher rate of death or HF rehospitalization in patients without SR restoration than patients with SR restoration (44% vs. 20%, p<0.05) (Figure A). Compared with 50 paired subjects from HIJ-HF 2 study, those who treated with landiolol developed a significant higher rate of SR restoration (68% vs. 20%, p<0.05) (Figure B).

Figure 1

Conclusion

This study demonstrated that landiolol treatment was effective for both rate control and conversion to sinus rhythm in ADHF patient with ATA. We should consider that the benefits of rhythm control in ADHF patients with ATA during and after landiolol treatment.

Effect of the CLE14 polypeptide on GLABRA2 homolog gene expression in rice and tomato roots

The CLAVATA3/ESR (CLE) plant polypeptides act as peptide hormones in various physiological and developmental aspects in a diverse array of land plants. One of the CLE family of genes, CLE14, is reported to induce root hair formation in Arabidopsis thaliana roots. Previously, we demonstrated that the application of synthetic CLE14 polypeptide treatment induced excess root hairs, and reduced the expression level of the non-hair cell fate determinant gene, GLABRA2 (GL2) in Arabidopsis roots. In this study, we investigated the function of synthetic CLE14 polypeptide in rice (Oryza sativa) and tomato (Solanum lycopersicum) roots. We measured the expression levels of the OsGL2 and SlGL2 genes, i.e., homologs of the Arabidopsis GL2 gene, in rice and tomato seedlings, respectively. Although CLE14 polypeptide treatment induced excess root hair formation in rice roots, substantial root hair induction was not observed in tomato roots. However, the CLE14 polypeptide treatment significantly inhibited the expression of the GL2 homolog genes of rice (OsGL2) and tomato (SlGL2). Our findings thus indicated that CLE14 can inhibit the GL2 gene expression in both rice and tomato plants, similar to the effect seen in Arabidopsis.

N-Ethyl-N′-(3-methylbenzoyl)-S,S-diphenylsulfodiimide

The asymmetric unit of the title sulfodiimide, C22H22N2OS, consists of two crystallographically independent molecules with similar conformations The environment around each sulfur atom is a slightly distorted tetrahedron with two S=N bonds and two S—C bonds. The S= N(m-methylbenzoyl) and S=N(NEt) bond lengths are 1.584 (3) and 1.528 (2) Å, respectively, for one molecule, and 1.575 (2) and 1.529 (3) Å, respectively, for the other. The dihedral angles between the two phenyl rings in the molecules are 86.76 (8) and 82.49 (8)°. The N—S—N—C(m-methylbenzoyl) and N—S—N—C(ethyl) torsion angles are −60.5 (2) and −50.28 (19)°, respectively, for one molecule, and 62.9 (2) and 44.2 (3)°, respectively, for the other. In the crystal, each independent molecule is linked to its inversion-related molecule via a pair of C—H...O hydrogen bonds, forming a dimer.

Association between chronic asymptomatic pancreatic hyperenzymemia and pancreatic ductal anomalies: a magnetic resonance cholangiopancreatography study

PURPOSE

Elucidating the association between pancreatic ductal anomalies and chronic asymptomatic pancreatic hyperenzymemia using magnetic resonance cholangiopancreatography.

METHODS

We conducted a single-center, retrospective, case-control study. The healthy community group comprised 554 subjects who participated in a paid, whole-body health checkup program. The patient group comprised 14 subjects with idiopathic pancreatic hyperamylasemia or hyperlipasemia. All subjects underwent magnetic resonance cholangiopancreatography. The clinical features and incidence rates of pancreatic ductal anomalies were then compared between the groups.

RESULTS

Compared to the healthy community group, the patient group was significantly more likely to be ≥ age 65 (71.4% of patient group vs. 22.1% of healthy community group), have a history of diabetes mellitus (21.4% vs. 5.4%) or hypertension (35.7% vs. 11.4%), and to have pancreas divisum (21.4% vs. 2.7%), meandering main pancreatic duct (21.4% vs. 4.1%), Wirsungocele (14.3% vs. 1.1%), or dilated main pancreatic duct (14.3% vs. 2.3%). Multivariate analysis found that age ≥ 65 (odds ratio 8.76), presence of pancreas divisum (odds ratio 13.2), meandering main pancreatic duct (odds ratio 8.95), and Wirsungocele (odds ratio 17.6) were independent factors significantly associated with chronic asymptomatic pancreatic hyperenzymemia.

CONCLUSIONS

Pancreas divisum, meandering main pancreatic duct, and Wirsungocele were independently associated with chronic asymptomatic pancreatic hyperenzymemia.

Abstract P1-11-01: Oral care evaluation to prevent oral mucositis in estrogen receptor positive metastatic breast cancer patients treated with everolimus (Oral Care-BC): A randomized controlled phase III trial

Introduction

Oral mucositis is a clinically significant complication of mucotoxic cancer therapy. The incidence of oral mucositis (any grade) as an adverse drug reaction of everolimus is 58%, while an analysis of Asian people has reported its occurrence as 81%.This study hypothesizes that the occurrence of oral mucositis will reduce with professional oral care (POC) administered prior to everolimus treatment.

Method:

This was a randomized, multi-center, open-label, phase III study, to evaluate the efficacy of POC in preventing mucositis induced by everolimus in postmenopausal, estrogen receptor (ER)-positive, metastatic breast cancer patients. Patients were randomized into POC and control groups (1:1 ratio). All patients received everolimus with exemestane and continued the everolimus until disease progression. In the POC group, patients were subjected to teeth surface cleaning, scaling and tongue cleaning, before initiating everolimus, and continued to receive weekly POC from dentist or oral surgeons throughout the 8 weeks of treatment. In the control group, patients brushed their own teeth and gargled with 0.9% sodium chloride solution or water. The primary end-point was to measure the incidence of all grades of oral mucositis. Target accrual was 200 patients with a 2-sided type I error rate of 5% and 80% power to detect 25% risk reduction. This trial has been registered at ClinicalTrials.gov, number NCT 02069093.

Result:

Between May 26, 2014 and Dec 28, 2017, we enrolled 174 women from 31 institutions; 168 were evaluable for efficacy but 5 were excluded (had not received the protocol treatment [n=4]; no efficacy data [n=1]). In 8 weeks, the incidence of grade 1 oral mucositis was significantly different between the POC group (76.5%, 62 of 81 patients) and control group (89.7%, 78 of 87 patients) (p=0.035). The incidence of grade 2 (severe) oral mucositis was also significantly different between the POC group (34.6%, 28 of 81 patients) and control group (54%, 47 of 87 patients) (p= 0.015). As a result of oral mucositis, 18 (22.2%) patients in the POC group and 28 (32.2%) in the control group had to undergo everolimus dose reduction.

Conclusion:

POC reduced the incidence and severity of oral mucositis in patients receiving everolimus and exemestane. This could be a new standard in oral care for patients undergoing this treatment.

Primary Analysis: Incidence Probability of Oral Mucositis POC Group (n=81) Controll (n=87)P-valuneOral Mucositis over Grade1n% n% Yes6276.5 7889.70.035No1923.5 910.3        Risk Difference, % (95% CI)-11.83 (-22.80, -0.85)    POC: Professional oral Care

Citation Format: Niikura N, Nakatukasa K, Amamiya T, Watanabe K-i, Hata H, Kikawa Y, Taniike N, Yamanaka T, Mitsunaga S, Nakagami K, Adachi M, Kondo N, Horii K, Hayashi N, Naito M, Kashiwabara K, Yamashita T, Umeda M, Mukai H, Ota Y. Oral care evaluation to prevent oral mucositis in estrogen receptor positive metastatic breast cancer patients treated with everolimus (Oral Care-BC): A randomized controlled phase III trial [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P1-11-01.

Abstract P2-08-59: Different predictive and prognostic impact of intra-tumor heterogeneity, tumor biology, and microenvironment in triple negative breast cancer

Background

Intra-tumor heterogeneity, tumor biology, and microenvironment in triple negative breast cancer (TNBC) has been reported to be predictive and prognostic factors. However, it is not well known how these factors are correlated each other according to response to chemotherapy and their prognosis. The aim of this study was to assess the predictive and prognostic impact of these factors in TNBC.

Method

Biopsy samples before neoadjuvant chemotherapy (NAC) from 59 TNBC patients who underwent surgery after NAC from 2001 to 2007 were retrospectively assessed. For tumor biology, tumors were classified as Hormonal related luminal androgen receptor (LAR) if &gt;10% staining of AR, Basal-like if positive for cytokeratin 5/6 and EGFR, and Others. Claudin 1 and p16 expression levels were assessed for intra-tumor heterogeneity. and stromal tumor-infiltrating lymphocytes (Str-Tils) levels for tumor-microenvironment were also assessed as low for ≤10%, Intermediate for 10-49%, and high for &gt;50%.The predictive and prognostic impact of clinicopathological factors including age, nuclear grade (NG), lymph node status, were also assessed. Overall survival (OS) and disease-free survival (DFS) were estimated using the Kaplan-Meier method and were compared between groups using the log-rank test.

Results

A median overall survival period of the 59 patients was 98 month(6 -172 month).Eighteen (30.5%) were classified in LAR, 16 (27.1%) in Basal-like, and 25 (42.4%) in Others. According to response to NAC, 10 patients (16.9%) achieved pathologic complete response (pCR). These biological classifications were not associated with pCR rate (p=0.135). high-p16 had significantly high pCR rate (p=0.046).However, Str-Tils level was not associated with pCR rate. Patients with lymph node metastasis had significantly low pCR rate (p=0.017).In terms of their prognosis, age&lt;50 had significantly shorter OS and DFS than that of age&gt;50 (OS: p=0.023, DFS: p=0.027). NG3 had a trend of short OS compared to NG 1 or 2 (NG 1 vs 3, OS: p=0.053, and NG 2 vs 3, OS: p=0.073). There were no difference of their prognosis among three tumor biology classifications except Basal-like had significantly shorter OS than that of LAR (LAR vs Basal OS:p=0.041, DFS:p=0.574, LAR vs Others OS:p=0.407, DFS:p=0.866, Basal vs Others OS:p=0.162, DFS:p=0.713).Claudin 1 and p16 expression levels were not associated with OS and DFS. Low-Str-Tils had a trend of shorter OS and DFS than that of intermediate- or high-Str-Tils (low vs int; OS:p=0.085, DFS:p=0.026, low vs high; OS:p=0.062, DFS:p=0.055).In multivariate analysis, age&lt;50 was only independent prognostic factor (p&lt;0.05).

Conclusion

We showed that intra-tumor heterogeneity, tumor biology, and microenvironment had different predictive and prognostic impact in TNBC. These results might suggest the strategy of additional targeting treatment to non-pCR patients.

Citation Format: Hayashi N, Nakamura M, Kobayashi D, Suzuki K, Yamauchi H. Different predictive and prognostic impact of intra-tumor heterogeneity, tumor biology, and microenvironment in triple negative breast cancer [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P2-08-59.

Abstract P2-14-17: Long-term oncologic safety of nipple-sparing mastectomy with immediate reconstruction

Background

Nipple-sparing mastectomy (NSM) is an alternative procedure to skin-sparing mastectomy (SSM) for selected patients who undergo immediate reconstruction. However, the evidence of long-term oncologic safety of NSM has not been established. In this study, we aimed to compare the prognosis of breast cancer patients who underwent NSM to those who underwent SSM with immediate reconstruction.

Methods

The clinicopathological factors including recurrence site, pathologic stage, nipple-tumor distance, histological type, lymphovascular invasion, margin status, ER, PgR and HER2 status of stage 0–III primary breast cancer patients who underwent NSM or SSM with immediate primary reconstruction with tissue expander from our breast center database was retrospectively assessed. Patients with a nipple tumor distance of &lt;1 cm who underwent NSM were excluded. 190 patients who underwent NSM and 729 patients who underwent SSM were included in the analysis. All patients underwent MRI or US before treatment. Nipple-tumor distance was mainly measured by MRI.

Results

The median follow-up period was 71 months (range: 10 - 131 months) for the NSM group and 79 months (range: 9 - 140 months) for the SSM group. There were no significant difference of clinicopathological factors between the NSM group and the SSM group, except of the larger diameter of tumor in the SSM group. NSM was performed for 60 patients (32%) with stage 0, 71 patients (37%) with stage I, and 59 patients (31%) with stage II/III. SSM was performed for 185 patients (26%) with stage 0, 268 patients (37%) with stage I, and 276 patients (37%) with stage II/III. Local recurrence was found in 11 (5.8%) patients in the NSM group and in 44 (6.0%) patients in the SSM group. In the NSM group, only one (0.5%) patient had local recurrence in the nipple areola complex. In terms of DFS and overall survival (OS) rate, there was no difference between the NSM group and the SSM group (DFS; 89.5% vs 89.2%, HR, 1.044; p = 0.8992; 95% CI, 0.5116–1.9519, and OS; 98.4% vs 96.4%, HR, 0.963; p = 0.9116; 95% CI, 0.473–1.793). According to breast cancer subtype, in the NSM group, all of the 11 patients (100%) who developed local recurrence in the NSM group was hormone receptor (HR)-positive/HER2-negative breast cancer. 29 of the 44 patients (65.9%) who developed local recurrence in the SSM group was HR-positive/HER2-negative, 6 patients (13.6%) was HR-negative/HER2-positive, and 7 patients (15.9%) was triple-negative breast cancer. Among patients who had received neoadjuvant chemotherapy, the NSM group (3 of 14 patients, 21.4%) had a trend for higher local recurrence rate than the SSM group 7 of 116 patients (6.0%) (p = 0.0813). However, no local recurrence in the nipple areola complex was observed for the NSM group. In addition, there was no difference of OS between the NSM group (92.9%) and the SSM group (90.5%) (HR, 0.903; p = 0.9943; 95% CI, 0.049-4.739).

Conclusions

Our results suggested that NSM with immediate reconstruction might be safe as well as SSM for breast cancer with the nipple–tumor distance of &gt;1 cm with respect to their prognosis and local control regardless of breast cancer subtype or invasiveness. Further studies with a large sample size to assess the risk of local recurrence for NSM after neoadjuvant chemotherapy.

Citation Format: Yamashita Y, Hayashi N, Nagura N, Kajiura Y, Yoshida A, Takei J, Suzuki K, Tsunoda H, Yamauchi H. Long-term oncologic safety of nipple-sparing mastectomy with immediate reconstruction [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P2-14-17.