Category: Nevin Manimala

Invest Radiol. 2021 May 19. doi: 10.1097/RLI.0000000000000791. Online ahead of print.

ABSTRACT

BACKGROUND: The potential of deep learning to support radiologist prostate magnetic resonance imaging (MRI) interpretation has been demonstrated.

PURPOSE: The aim of this study was to evaluate the effects of increased and diversified training data (TD) on deep learning performance for detection and segmentation of clinically significant prostate cancer-suspicious lesions.

MATERIALS AND METHODS: In this retrospective study, biparametric (T2-weighted and diffusion-weighted) prostate MRI acquired with multiple 1.5-T and 3.0-T MRI scanners in consecutive men was used for training and testing of prostate segmentation and lesion detection networks. Ground truth was the combination of targeted and extended systematic MRI-transrectal ultrasound fusion biopsies, with significant prostate cancer defined as International Society of Urological Pathology grade group greater than or equal to 2. U-Nets were internally validated on full, reduced, and PROSTATEx-enhanced training sets and subsequently externally validated on the institutional test set and the PROSTATEx test set. U-Net segmentation was calibrated to clinically desired levels in cross-validation, and test performance was subsequently compared using sensitivities, specificities, predictive values, and Dice coefficient.

RESULTS: One thousand four hundred eighty-eight institutional examinations (median age, 64 years; interquartile range, 58-70 years) were temporally split into training (2014-2017, 806 examinations, supplemented by 204 PROSTATEx examinations) and test (2018-2020, 682 examinations) sets. In the test set, Prostate Imaging-Reporting and Data System (PI-RADS) cutoffs greater than or equal to 3 and greater than or equal to 4 on a per-patient basis had sensitivity of 97% (241/249) and 90% (223/249) at specificity of 19% (82/433) and 56% (242/433), respectively. The full U-Net had corresponding sensitivity of 97% (241/249) and 88% (219/249) with specificity of 20% (86/433) and 59% (254/433), not statistically different from PI-RADS (P > 0.3 for all comparisons). U-Net trained using a reduced set of 171 consecutive examinations achieved inferior performance (P < 0.001). PROSTATEx training enhancement did not improve performance. Dice coefficients were 0.90 for prostate and 0.42/0.53 for MRI lesion segmentation at PI-RADS category 3/4 equivalents.

CONCLUSIONS: In a large institutional test set, U-Net confirms similar performance to clinical PI-RADS assessment and benefits from more TD, with neither institutional nor PROSTATEx performance improved by adding multiscanner or bi-institutional TD.

PMID:34049336 | DOI:10.1097/RLI.0000000000000791

J Nurs Res. 2021 May 27. doi: 10.1097/JNR.0000000000000437. Online ahead of print.

ABSTRACT

BACKGROUND: Urinary catheters (UCs) with a closed urinary drainage system have been widely used in patients for many years. However, the frequency of replacing and operating these devices may be associated with catheter-associated urinary tract infection (CAUTI).

PURPOSE: This study was designed to compare the incidence of CAUTI by replacement time (every 14 or ≥ 15 days) of the urinary drainage system.

METHODS: This 1-year prospective, nonrandomized controlled study was conducted in a major teaching hospital. The Transparent Reporting of Evaluations with Nonrandomized Designs Statement checklist was used. All of the patients with UCs were divided into two groups based on each patient’s preference with regard to replacement time of the urinary drainage system.

RESULTS: Five hundred sixty-two patients were evaluated, and 341 patients with UCs were enrolled as participants in the study. In the per-protocol analysis, 16 patients (22.2%; 9.3 episodes/1,000 catheter-days) in the 14-day group and 15 patients (17.9%; relative risk = 1.24, 95% confidence interval [0.66, 2.34]) in the ≥ 15-day group (7.7 episodes/1,000 catheter-days; incidence density ratio 1.20, 95% confidence interval [0.60, 2.43]) had CAUTIs. A comparison of cleanliness within urinary bags showed no significant intergroup difference (p > .05). In the intention-to-treat analysis, the incidence of CAUTI between the two groups was also not significantly different (p > .05).

CONCLUSIONS: No statistically significant difference in the incidence of CAUTI was identified between patients who used the 14-day replacement interval and those who used the ≥ 15-day replacement interval for their urinary drainage system.

PMID:34049325 | DOI:10.1097/JNR.0000000000000437

Otol Neurotol. 2021 May 26. doi: 10.1097/MAO.0000000000003204. Online ahead of print.

ABSTRACT

OBJECTIVE: To examine if performing stapedotomy as the first case of the day provides improved outcomes compared with those performed later in the day.

STUDY DESIGN: Retrospective chart review.

SETTING: Tertiary referral center.

PATIENTS: Adult patients undergoing stapedotomy for otosclerosis.

MAIN OUTCOME MEASURES: Patients were separated into either a first case group or a later case group based on surgical start time. Audiologic outcomes and complications were compared between the two groups.

RESULTS: The first case group had a smaller postoperative air-bone gap (ABG) compared with the later case group of 9.81 dB HL compared with 11.73dB HL and 3.79 dB HL compared with 6.29 dB HL at 1000 and 2000 Hz, respectively (p = 0.03, p < 0.01). The mean postoperative ABG was 10.63 dB HL for the first start group compared with 12.12 dB HL for the later start group, which was statistically significant (p = 0.05).

CONCLUSIONS: First start stapedotomy is associated with slightly improved audiologic outcomes compared with those starting later in the day, although both groups had significantly improved postoperative outcomes overall. There was no significant difference in complications when comparing stapedotomy by case start time.

PMID:34049326 | DOI:10.1097/MAO.0000000000003204

Otol Neurotol. 2021 May 26. doi: 10.1097/MAO.0000000000003200. Online ahead of print.

ABSTRACT

OBJECTIVE: Characterize the speech recognition and sound source localization of patients with unilateral Meniere’s disease who undergo labyrinthectomy for vertigo control with simultaneous or sequential cochlear implantation.

DATABASES REVIEWED: PubMed, Embase, and Cochrane databases.

METHODS: The search was performed on May 6, 2020. The keywords utilized included: “Meniere’s disease AND cochlear implant”; “cochlear implant AND single sided deafness”; “cochlear implant AND vestibular”; and “labyrinthectomy AND cochlear implant”. Manuscripts published in English with a publication date after 1995 that assessed adult subjects (≥18 years of age) were included for review. Subjects must have been diagnosed with Meniere’s disease unilaterally and underwent labyrinthectomy with simultaneous or sequential cochlear implantation. Reported outcomes with cochlear implant (CI) use included speech recognition as measured with the Consonant-Nucleus-Consonant (CNC) word test and/or sound source localization reported in root-mean squared (RMS) error. The method of data collection and study type were recorded to assess level of evidence. Statistical analysis was performed with Wilcoxon signed ranks test.

RESULTS: Data from 14 CI recipients met the criteria for inclusion. Word recognition comparisons between the pre-operative interval and a post-activation interval demonstrated a significant improvement with the CI (p = 0.014), with an average improvement of 23% (range -16-50%). Sound source localization post-operatively with the CI demonstrated an average RMS error of 26° (SD 6.8, range 18.7-43.1°) compared to the 42° (SD 19.1, range 18-85°) in the pre-operative or CI off condition, these two conditions were not statistically different (p = 0.148).

CONCLUSION: Cochlear implantation and labyrinthectomy in adult patients with Meniere’s disease can support improvements in speech recognition and sound source localization for some CI users, though observed performance may be poorer than traditional CI candidates.

PMID:34049331 | DOI:10.1097/MAO.0000000000003200

ASAIO J. 2021 May 28. doi: 10.1097/MAT.0000000000001490. Online ahead of print.

ABSTRACT

Maintaining dialysis vascular access is a source of considerable morbidity in patients with end-stage renal disease (ESRD). High-resolution radiofrequency (RF) ultrasound vascular strain imaging has been applied experimentally in the vascular access setting to assist in diagnosis and management. Unfortunately, high-resolution RF data are not routinely accessible to clinicians. In contrast, the standard DICOM formatted B-mode ultrasound data are widely accessible. However, B-mode, representing the envelope of the RF signal, is of much lower resolution. If strain imaging could use open-source B-mode data, these imaging techniques could be more broadly investigated. We conducted experiments to detect wall strain signals with submillimeter tracking resolutions ranging from 0.2 mm (3 pixels) to 0.65 mm (10 pixels) using DICOM B-mode data. We compared this submillimeter tracking to the overall vascular distensibility as the reference measurements to see if high-strain resolution strain could be detected using open-source B-Mode data. We measured the best-fit coefficient of determination between signals, expressed as the percentage of strain waveforms that exhibited a correlation with a p value of 0.05 or less. The lowest percentage was 86.7%, and most were 90% and higher. This indicates high-resolution strain signals can be detected within the vessel wall using B-mode DICOM data.

PMID:34049311 | DOI:10.1097/MAT.0000000000001490

J Nurs Res. 2021 May 27. doi: 10.1097/JNR.0000000000000435. Online ahead of print.

ABSTRACT

BACKGROUND: The World Health Organization encourages countries to improve birth outcomes to reduce rates of neonatal mortality and morbidity.

PURPOSE: This study was designed to examine the effect of environmental crude oil pollution on newborn birth outcomes in Rivers State, Nigeria.

METHODS: A retrospective cohort design was used to examine the effects of exposure to oil pollution on birth outcomes using facility-based records. K-Dere (an oil-polluted community) served as the exposure group, whereas birth records from Iriebe served as the comparison group. A sample size of 338 systematically selected birth records was examined (169 records for each arm of the study). A data extraction sheet was used for data collection. Data were analyzed using descriptive and inferential statistics at p < .05.

RESULTS: The risk of preterm birth was significantly higher in the exposed group (16% vs. 7.7%, relative risk = 2.08, 95% CI [1.11, 3.89], p = .018). At 6 weeks after birth, newborns in the exposed group weighed significantly less (4.64 ± 0.82 vs. 4.85 ± 0.92 kg, p = .032) and reported significantly higher incidence of morbidity compared with the newborns in the comparison group (relative risk = 3.03, 95% CI [2.20, 4.19], p < .001).

CONCLUSIONS: The oil-polluted area examined in this study was found to have a higher risk of preterm birth, a slower rate of newborn growth, and a higher rate of newborn morbidity than the non-oil-polluted area at 6 weeks after birth. Stakeholders should sustain efforts to remediate the environment in polluted regions and prevent oil pollution. Future research should investigate the mechanisms of the observed toxicological effects and the targeted protection of vulnerable groups in oil-polluted communities.

PMID:34049324 | DOI:10.1097/JNR.0000000000000435

Phys Med Biol. 2021 May 28. doi: 10.1088/1361-6560/ac0683. Online ahead of print.

ABSTRACT

Multiple injection dynamic positron emission tomography (PET) scanning is used in the clinical management of certain groups of patients and in medical research. The analysis of these studies can be approached in two ways: (i) separate analysis of data from individual tracer injections, or (ii), concatenate/pool data from separate injections and carry out a combined analysis. The simplicity of separate analysis has some practical appeal but may not be statistically efficient. We use a linear model framework associated with a kinetic mapping scheme to develop a simplified theoretical understanding of separate and combined analysis. The theoretical framework is explored numerically using both 1-D and 2-D simulation models. These studies are motivated by the breast cancer flow-metabolism mismatch studies involving 15O-Water (H2O) and 18F-Fluorodeoxyglucose (FDG) and repeat 15O-H2O injections used in brain activation investigations. Numerical results are found to be substantially in line with the simple theoretical analysis: mean square error (MSE) characteristics of alternative methods are well described by factors involving the local voxel-level resolution of the imaging data, the relative activities of the individual scans and the number of separate injections involved. While voxel-level resolution has dependence on scan dose, after adjustment for this effect, the impact of a combined analysis is understood in simple terms associated with the linear model used for kinetic mapping. This is true for both data reconstructed by direct filtered backprojection (FBP) or iterative maximum likelihood (ML). The proposed analysis has potential to be applied to the emerging long axial field-of-view PET scanners.

PMID:34049293 | DOI:10.1088/1361-6560/ac0683

Phys Med Biol. 2021 May 28. doi: 10.1088/1361-6560/ac0684. Online ahead of print.

ABSTRACT

CT images for radiotherapy planning are usually acquired in thick slice to reduce imaging dose, especially for pediatric patients, and to lessen the need for contouring and treatment planning on more slices. However, low through-plane resolution may degrade the accuracy of dose calculations. In this paper, a self-supervised deep learning workflow is proposed to synthesize high through-plane resolution CT images by learning from their high in-plane resolution features. The proposed workflow was designed to facilitate the neural networks to learn the mapping from low resolution (LR) to high resolution (HR) images in the axial plane. During the inference step, the HR sagittal and coronal images were generated by feeding two parallelly trained neural networks with the respective LR sagittal and coronal images to the trained neural networks. The CT simulation images of a cohort of 75 head and neck (HN) cancer patients (1 mm slice thickness) and 200 CT images of a cohort of 20 lung cancer patients (3 mm slice thickness) were retrospectively investigated with a cross validation manner. The generated HR images with the proposed method were qualitatively (visual quality, image intensity profiles and preliminary observer study) and quantitatively (Mean Absolute Error (MAE), Edge Keeping Index (EKI), Structural Similarity Index Measurement (SSIM), Information Fidelity Criterion (IFC) and Visual Information Fidelity in Pixel domain (VIFP)) inspected, while taking the original HN and lung cancer patients’ CT images as the reference. The qualitative results have shown the capability of the proposed method for generating high through-plane resolution CT images with data of the HN and lung cancer patients. All the improvements of the measure metrics are confirmed to be statistically significant with paired two-sample t-test analysis. The innovative point of the work is that the proposed deep learning workflow for CT image generation with high through-plane resolution in radiotherapy is self-supervised, which means it does not rely on ground truth CT images to train the network. In addition, the assumption that the in-plane HR information can supervise the through-plane HR generation is confirmed and anticipated to potentially inspire more researches on this topic to further improve the through-plane resolution of medical images.

PMID:34049297 | DOI:10.1088/1361-6560/ac0684

Phys Med Biol. 2021 May 28. doi: 10.1088/1361-6560/ac0681. Online ahead of print.

ABSTRACT

For decades, dose-volume information for segmented anatomy has provided the essential data for correlating radiotherapy dosimetry with treatment-induced complications. Dose-volume information has formed the basis for modelling those associations via normal tissue complication (NTCP) models and for driving treatment planning. Limitations to this approach have been identified. Many studies have emerged demonstrating that the incorporation of information describing the spatial nature of the dose distribution, and potentially its correlation with anatomy, can provide more robust associations with toxicity and seed more general NTCP models. Such approaches are culminating in the application of computationally intensive processes such as machine learning and the application of neural networks. The opportunities these approaches have for individualising treatment, predicting toxicity and expanding the solution space for radiation therapy are substantial and have clearly widespread and disruptive potential. Impediments to reaching that potential include issues associated with data collection, model generalisation and validation. This review examines the role of spatial models of complication and summarises relevant published studies. Sources of data for these studies, appropriate statistical methodology, frameworks for processing spatial dose information and extracting relevant features are described. Spatial complication modelling is consolidated as a pathway to guiding future developments towards effective, complication-free radiotherapy treatment.

PMID:34049304 | DOI:10.1088/1361-6560/ac0681

Appl Ergon. 2021 May 25;96:103462. doi: 10.1016/j.apergo.2021.103462. Online ahead of print.

ABSTRACT

Hand anthropometry is one of the fundamentals of ergonomic research and product design. Many studies have been conducted to analyze the hand dimensions among different populations, however, the definitions and the numbers of those dimensions were usually selected based on the experience of the researchers and the available equipment. Few studies explored the importance of each hand dimension regarding the 3D shape of the hand. In this paper, we aim to identify the dominant dimensions that influence the hand shape variability while considering the stability of the measurements in practice. A novel four-step research method was proposed where in the first step, based on literature study, we defined 58 landmarks and 53 dimensions for the exploration. In the second step, 80,000 virtual hand models, each had the associated 53 dimensions, were augmented by changing the weights of Principle Components (PCs) of a statistical shape model (SSM). Deep neural networks (DNNs) were used to establish the inverse relationships from the dimensions to the weight of each PC of the hand SSM. Using the structured sparsity learning method, we identified 21 dominant dimensions that represent 90% of the variance of the hand shape. In the third step, two different manual measuring methods were used to evaluate the stability of the measurements in practice. Finally, we selected 16 dominant dimensions with lower measurement variance by synthesizing the findings in Step 2 and 3. It was concluded that the recognized 21 dominant dimensions can be treated as the reference dimensions for anthropometric study and using the selected 16 dominant dimensions with lower measurement variance, ergonomists are able to generate a 3D hand model based on simple measurement tools with an accuracy of 5.9 mm. Though the accuracy is limited, the efforts are minimum, and the results can be used as an indicator in the early stage of research/design.

PMID:34049195 | DOI:10.1016/j.apergo.2021.103462