Convolutional neural network -based phantom image scoring for mammography quality control.

BACKGROUND: Visual evaluation of phantom images is an important, but time-consuming part of mammography quality control (QC). Consistent scoring of phantom images over the device's lifetime is highly desirable. Recently, convolutional neural networks (CNNs) have been applied to a wide range of image classification problems, performing with a high accuracy. The purpose of this study was to automate mammography QC phantom scoring task by training CNN models to mimic a human reviewer. METHODS: Eight CNN variations consisting of three to ten convolutional layers were trained for detecting targets (fibres, microcalcifications and masses) in American College of Radiology (ACR) accreditation phantom images and the results were compared with human scoring. Regular and artificially degraded/improved QC phantom images from eight mammography devices were visually evaluated by one reviewer. These images were used in training the CNN models. A separate test set consisted of daily QC images from the eight devices and separately acquired images with varying dose levels. These were scored by four reviewers and considered the ground truth for CNN performance testing. RESULTS: Although hyper-parameter search space was limited, an optimal network depth after which additional layers resulted in decreased accuracy was identified. The highest scoring accuracy (95%) was achieved with the CNN consisting of six convolutional layers. The highest deviation between the CNN and the reviewers was found at lowest dose levels. No significant difference emerged between the visual reviews and CNN results except in case of smallest masses. CONCLUSION: A CNN-based automatic mammography QC phantom scoring system can score phantom images in a good agreement with human reviewers, and can therefore be of benefit in mammography QC.

Routine Ultrasound Quality Assurance in a Multi-Unit Radiology Department: A Retrospective Evaluation of Transducer Failures.

The importance of quality assurance (QA) in medical ultrasound (US) has been widely recognized and recommendations concerning technical QA have been published over the years. However, the demonstrated impact of a properly working QA protocol on clinical routine has been scarce. We investigated the transducer write-off causes for a 5-y period in a multi-unit radiology department with an annual average of 230 transducers in demanding diagnostic use. The transducer faults and the initial observers of the faults leading to transducer write-offs were traced and categorized. The most common cause of transducer write-off was an image uniformity problem or element failure. Mechanical faults or excessive leakage current and defects in the lens constituted smaller yet substantial shares. Our results suggest that a properly working routine QA program can detect majority of the faults before they are reported by users.

Protocol for motor and language mapping by navigated TMS in patients and healthy volunteers; workshop report.

INTRODUCTION: Navigated transcranial magnetic stimulation (nTMS) is increasingly used for preoperative mapping of motor function, and clinical evidence for its benefit for brain tumor patients is accumulating. In respect to language mapping with repetitive nTMS, literature reports have yielded variable results, and it is currently not routinely performed for presurgical language localization. The aim of this project is to define a common protocol for nTMS motor and language mapping to standardize its neurosurgical application and increase its clinical value. METHODS: The nTMS workshop group, consisting of highly experienced nTMS users with experience of more than 1500 preoperative nTMS examinations, met in Helsinki in January 2016 for thorough discussions of current evidence and personal experiences with the goal to recommend a standardized protocol for neurosurgical applications. RESULTS: nTMS motor mapping is a reliable and clinically validated tool to identify functional areas belonging to both normal and lesioned primary motor cortex. In contrast, this is less clear for language-eloquent cortical areas identified by nTMS. The user group agreed on a core protocol, which enables comparison of results between centers and has an excellent safety profile. Recommendations for nTMS motor and language mapping protocols and their optimal clinical integration are presented here. CONCLUSION: At present, the expert panel recommends nTMS motor mapping in routine neurosurgical practice, as it has a sufficient level of evidence supporting its reliability. The panel recommends that nTMS language mapping be used in the framework of clinical studies to continue refinement of its protocol and increase reliability.

Accelerometer-based automatic voice onset detection in speech mapping with navigated repetitive transcranial magnetic stimulation.

BACKGROUND: The use of navigated repetitive transcranial magnetic stimulation (rTMS) in mapping of speech-related brain areas has recently shown to be useful in preoperative workflow of epilepsy and tumor patients. However, substantial inter- and intraobserver variability and non-optimal replicability of the rTMS results have been reported, and a need for additional development of the methodology is recognized. In TMS motor cortex mappings the evoked responses can be quantitatively monitored by electromyographic recordings; however, no such easily available setup exists for speech mappings. NEW METHOD: We present an accelerometer-based setup for detection of vocalization-related larynx vibrations combined with an automatic routine for voice onset detection for rTMS speech mapping applying naming. COMPARISON WITH EXISTING METHOD(S): The results produced by the automatic routine were compared with the manually reviewed video-recordings. RESULTS: The new method was applied in the routine navigated rTMS speech mapping for 12 consecutive patients during preoperative workup for epilepsy or tumor surgery. The automatic routine correctly detected 96% of the voice onsets, resulting in 96% sensitivity and 71% specificity. Majority (63%) of the misdetections were related to visible throat movements, extra voices before the response, or delayed naming of the previous stimuli. The no-response errors were correctly detected in 88% of events. CONCLUSION: The proposed setup for automatic detection of voice onsets provides quantitative additional data for analysis of the rTMS-induced speech response modifications. The objectively defined speech response latencies increase the repeatability, reliability and stratification of the rTMS results.

Integrating nTMS Data into a Radiology Picture Archiving System.

Navigated transcranial magnetic stimulation (nTMS) is employed in eloquent brain area localization prior to intraoperative direct cortical electrical stimulations and neurosurgery. No commercial archiving or file transfer protocol existed for these studies. The aim of our project was to establish a standardized protocol for the transfer of nTMS results and medical assessments to the end users in pursuance of improving data security and facilitating presurgical planning. The existing infrastructure of the hospital's Radiology Department was used. Hospital information systems and networks were configured to allow communications and archiving of the study results, and in-house software was written for file manipulations and transfers. Graphical user interface with description suggestions and user-defined text legends enabled an easy and straightforward workflow for annotations and archiving of the results. The software and configurations were implemented and have been applied in studies of ten patients. The creation of the study protocol required the involvement of various professionals and interdepartmental cooperation. The introduction of the protocol has ended previously recurrent involvement of staff in the file transfer phase and improved cost-effectiveness.

Applicability of nTMS in locating the motor cortical representation areas in patients with epilepsy.

BACKGROUND: Transcranial magnetic stimulation (TMS) is increasingly used for non-invasive functional mapping in preoperative evaluation for brain surgery, and the reliability of navigated TMS (nTMS) motor representation maps has been studied in the healthy population and in brain tumor patients. The lesions behind intractable epilepsy differ from typical brain tumors, ranging from developmental cortical malformations to injuries early in development, and may influence the functional organization of the cortical areas. Moreover, the interictal cortical epileptic activity and antiepileptic medication may affect the nTMS motor threshold. The reliability of the nTMS motor representation localization in epilepsy patients has not been addressed. METHODS: We compared the nTMS motor cortical representation maps of hand and arm muscles with the results of invasive electrical cortical stimulation (ECS) in 13 patients with focal epilepsy. The nTMS maps were projected to the cortical surface segmented from preoperative magnetic resonance images (MRI), and the positions of the subdural electrodes were extracted from the postoperative low-dose computed tomography (CT) images registered with preoperative MRI. RESULTS: The 3D distance between the average nTMS site and average ECS electrode location was 11 ± 4 mm for the hand and 16 ± 7 mm for arm muscle representation areas. In all patients the representation areas defined with nTMS and ECS were located on the same gyrus, also in patients with abundant interictal epileptic activity on the motor gyrus. CONCLUSIONS: nTMS can reliably locate the hand motor cortical representation area with the accuracy needed for pre-surgical evaluation in patients with epilepsy.

Functional plasticity of the motor cortical structures demonstrated by navigated TMS in two patients with epilepsy.

BACKGROUND: Recently, navigated transcranial magnetic stimulation (nTMS) has been suggested to be useful in preoperative functional localization of motor cortex in patients having tumors close to the somatomotor cortex. Resection of tumors in anatomically predicted eloquent areas without adverse effects have emphasized functional plasticity elicited by intracranial pathology. OBJECTIVE: To describe functional plasticity of motor cortex indicated by nTMS in two patients with epilepsy. METHODS: nTMS, functional MRI (fMRI), diffusion-tensor (DT)-tractography and magnetoencephalography (MEG) were utilized to preoperatively localize motor cortical areas in the workup for epilepsy surgery. The localizations were compared with each other, with the cortical anatomical landmarks, and in one patient with invasive electrical cortical stimulation (ECS). RESULTS: In two out of 19 studied patients, nTMS identified motor cortical sites that differed from those indicated by anatomical landmarks. In one patient, nTMS activated preferentially premotor cortex rather than pathways originating from the precentral gyrus. MEG and fMRI localizations conformed with nTMS whereas ECS localized finger motor function into the precentral gyrus. Resection of the area producing motor responses in biphasic nTMS did not produce a motor deficit. In the other patient, nTMS indicated abnormal ipsilateral hand motor cortex localization and confirmed the functionality of aberrant motor cortical representations of the left foot also indicated by fMRI and DT-tractography. CONCLUSION: nTMS may reveal the functional plasticity and shifts of motor cortical function. Epileptic foci may modify cortical inhibition and the nTMS results. Therefore, in some patients with epilepsy, the nTMS results need to be interpreted with caution with regard to surgical planning.

Probing modifications of cortical excitability during stroke recovery with navigated transcranial magnetic stimulation.

OBJECTIVE: To follow cortical excitability changes during recovery from stroke with navigated transcranial magnetic stimulation (nTMS), in particular, to characterize changes of short-interval intracortical inhibition (SICI) and intracortical facilitation (ICF), to correlate them with recovery of upper extremity function, and to detect possible shifts of cortical hand representations. METHODS: Single and paired pulse nTMS were delivered to the hemisphere with infarction and to the hemisphere without infarction in 14 first-ever stroke patients at 1 (T1) and 3 months (T2) after stroke. Electromyographic responses to nTMS stimulation were recorded from the first dorsal interosseus muscles. nTMS was used to ensure an accurate coil repositioning in repeated measurements. Hand function recovery was clinically evaluated using the Action Research Arm Test (ARAT) and 9-hole peg test (9-HPT). RESULTS: SICI and ICF were modulated in both hemispheres during recovery. Inhibition in the hemisphere without infarction correlated significantly with the affected hand performance at T2; stronger disinhibition (poor inhibition) was associated with worse hand performance. Location of hand muscle representations was shifted in 3 well-recovered patients out of 14 patients at T2. CONCLUSIONS: In line with earlier studies, disinhibition in the hemisphere without infarction may be related to poor recovery of the affected hand. Usage of the affected hand during stroke recovery seems to influence these cortical excitability changes. nTMS is a valuable tool for tracking muscle cortical representation changes during brain reorganization.

Morphological analysis of sinus and retrograde atrial waves detected through a permanent pacemaker atrial lead.

AIMS: The main objective of this study was to characterize morphological differences between sinus and retrograde atrial waves. METHODS: We collected intracardiac atrial signals through a DDDR pacemaker to characterize their morphologies and discriminate retrograde from sinus atrial waves off-line. Intracardiac unipolar and bipolar signals were collected at an 800-Hz sampling rate through a 0.4-Hz high-pass filter. Sinus and retrograde atrial waves during ventricular pacing were recorded in the supine and upright positions. RESULTS: Eleven different form parameters (FPs) were applied to describe atrial wave morphology. Data from 14 patients were collected and analyzed. Atrial signals differed markedly between the two body postures. However, it was possible to discriminate retrograde from sinus atrial waves on the basis of at least one FP in 12 of 14 (86%) patients when unipolar and bipolar atrial signals recorded in both body postures were analyzed separately. When body postures were pooled together to mimic natural conditions, discrimination was successful in nine of 14 patients with either configuration of the atrial signal. CONCLUSIONS: The results indicate that retrograde atrial waves can be discriminated from sinus waves by using high sampling rate, digital signal processing, and specific FPs incorporated in these pacemakers.