Differentiating viral and bacterial infections: A machine learning model based on routine blood test values.

The growing threat of antibiotic resistance necessitates accurate differentiation between bacterial and viral infections for proper antibiotic administration. In this study, a Virus vs. Bacteria machine learning model was developed to distinguish between these infection types using 16 routine blood test results, C-reactive protein concentration (CRP), biological sex, and age. With a dataset of 44,120 cases from a single medical center, the model achieved an accuracy of 82.2 %, a sensitivity of 79.7 %, a specificity of 84.5 %, a Brier score of 0.129, and an area under the ROC curve (AUC) of 0.905, outperforming a CRP-based decision rule. Notably, the machine learning model enhanced accuracy within the CRP range of 10-40 mg/L, a range where CRP alone is less informative. These results highlight the advantage of integrating multiple blood parameters in diagnostics. The "Virus vs. Bacteria" model paves the way for advanced diagnostic tools, leveraging machine learning to optimize infection management.

Clinical and neurophysiological recovery of ulnar nerve conduction block at the elbow.

INTRODUCTION/AIMS: An important mechanism of peripheral nerve motor and sensory dysfunction is conduction block (CB). However, recovery from mechanically induced CB has been rarely studied in humans. The aim of this study was to describe clinical, electrodiagnostic (EDx), and ultrasonographic (US) characteristics of CB recovery in ulnar neuropathy at the elbow (UNE). METHODS: We recruited a group of consecutive patients presenting to our EDx laboratory with UNE and >50% motor CB. Patients' histories were obtained and neurologic, EDx, and US examinations were repeated every 1-3 mo for at least 12 mo. RESULTS: We studied 10 patients (5 men), with a mean age of 63 y (range, 51-81 y). In all affected arms CB was localized to the retrocondylar groove. Following conservative management, myometrically measured index finger abduction improved from a median of 49% to 100% relative to the contralateral index finger, and ulnar nerve CB decreased from a median of 74% to 6%. Most of the improvement took place within 8 mo of symptom onset, and 6 mo after receiving treatment instructions. Mean motor nerve conduction velocity improved from 15 to 27 m/s in the most affected 2-cm ulnar nerve segment. DISCUSSION: The resolution of CB after typical chronic compression may take longer than after acute compression. This should be considered by clinicians when estimating prognosis for discussions with patients.

No Major Nerve Regeneration Seems to Occur during Recovery of Ulnar Neuropathy at the Elbow.

Introduction: There are three main potential mechanisms of recovery after nerve lesion: (1) resolution of conduction block, (2) collateral reinnervation, and (3) nerve regeneration. Their relative contributions in recovery after focal neuropathies are not well established. Methods: In a group of previously reported prospective cohort of patients with ulnar neuropathy at the elbow (UNE), I performed a post-hoc analysis of their clinical and electrodiagnostic findings. I compared amplitudes of the compound muscle action potential (CMAP) and sensory nerve action potential (SNAP) on ulnar nerve stimulation, as well as qualitative concentric needle electromyography (EMG) findings in the abductor digiti minimi muscle on the initial and follow-up examinations several years later. Results: Altogether, 111 UNE patients (114 arms) were studied. During median follow-up period of 880 days (range: 385-1545 days), CMAP amplitude increased (p = 0.02), and conduction block in the elbow segment recovered (from median 17% to 7%; p < 0.001). By contrast, SNAP amplitude did not change (p = 0.89). On needle EMG, spontaneous denervation activity diminished (p < 0.001), motor unit potential (MUP) amplitude increased (p < 0.001), and MUP recruitment remained unchanged (p = 0.43). Conclusions: Findings of the present study indicate that nerve function in chronic focal compression/entrapment neuropathies seems to improve mainly due to the resolution of the conduction block and collateral reinnervation. Contribution of nerve regeneration seems to be minor; the majority of axons lost in chronic focal neuropathies probably never recover. Further studies using quantitative methods are needed to validate present findings.

Ultrasonography of peripheral nerve tumours: a case series.

BACKGROUND: Peripheral nerve tumours (PNTs) are rare, but important cause of peripheral nerve dysfunction. The aim of the study was to present a series of consecutive patients with PNTs evaluated in authors' ultrasonography (US) practice. PATIENTS AND METHODS: The electronic medical records of patients with PNTs examined at our US laboratory from February 2013 to May 2020 were retrospectively reviewed. Data on gender, age, clinical features, PNT location, electrodiagnostic (EDx) features and US findings were collected. RESULTS: In the analyzed period 2845 patients were examined in our US laboratory. From these 15 patients (0.5%) with PNTs were identified. Four of them (3 with confirmed neurofibromatosis) had multiple PNTs. Half of patients (53%) presented with features of peripheral nerve damage, and others with palpable mass or pain. The most often involved nerve was ulnar (36%). PNT cross sectional areas varied from 24 mm2 to 1250 mm2 (median, 61 mm2). Based in 5 patients on histological and in remaining patients on US features, schwannoma was diagnosed in 40%, neurofibroma in 27%, and perineurioma in 27% of patients. CONCLUSIONS: As in previous reports, PNTs in our series presented with neurological symptoms, palpable mass or pain. In contrast to other focal neuropathies, particularly nerves with schwannomas, in spite of their large thickening, often demonstrated well preserved function. Adding US to our clinical practice, enabled us to diagnose these rare peripheral nerve lesions that we missed before.

Expert consensus on the combined investigation of carpal tunnel syndrome with electrodiagnostic tests and neuromuscular ultrasound.

Expert consensus was sought to guide clinicians on the use of electrodiagnostic tests (EDX) and neuromuscular ultrasound (NMUS) in the investigation of suspected carpal tunnel syndrome (CTS). Consensus was achieved using the Delphi method via three consecutive anonymised surveys of 15 experts and was defined as rating agreement ≥ 80%. The panel agreed that combining EDX and NMUS is more informative than using each modality alone. NMUS adds value in patients with clinically suspected CTS with non-localizing or normal EDX, atypical EDX, failed CTS surgery, polyneuropathy, and CTS suspected to be secondary to structural pathology. The median nerve cross-sectional area should be measured at the site of maximal nerve enlargement, and the nerve should be scanned from mid-forearm to the palm. The group also identified those situations where the wrist-to-forearm area ratio and longitudinal scans of the median nerve should also be obtained. EDX should always be performed to quantify CTS severity and in individuals over age 70. This document is an initial step to guide clinicians on the combined investigation of CTS using EDX and NMUS, to be updated regularly with the emergence of new research.

Length of affected nerve segment in ulnar neuropathies at the elbow.

OBJECTIVE: To establish length of the affected nerve segment (LANS) in ulnar neuropathy at the elbow (UNE). METHODS: In a group of our previously reported UNE patients we identified 2-cm segments with reduced motor nerve conduction velocity (MNCV) on electrodiagnostic (EDx) studies and increased nerve cross-sectional areas (CSA) on ultrasonographic (US) studies. LANS was obtained by summation of these abnormal 2-cm segments separately for each approach. We also studied effect of selected independent parameters on LANS. RESULTS: Altogether we studied 189 patients (194 arms). Mean (SD) LANS determined in 171 arms with reduced ulnar MNCV was 4.15 (1.89) cm, and was similar (p = 0. 21) to LANS obtained in 147 arms with increased CSA 4.46 (2.29) cm. Longer LANS were found in right arms, clinically severe UNE, axonal UNE and UNE due to entrapment. The most commonly affected 6 cm segment included 89% of abnormal 2-cm segments, with 50% of included 2-cm segments being normal. By contrast, the whole 10 cm segment included all abnormal 2-cm segments, with 66% of included segments being normal. CONCLUSIONS: In UNE both EDx and US studies revealed average LANS of around 4 cm. LANS was longer in more severe UNE. SIGNIFICANCE: LANS needs to be taken into account in discussion of the mechanisms of UNE and approach to EDx diagnosis of UNE, particularly length of the segment used in nerve conduction studies across the elbow.

Expert consensus on the combined investigation of ulnar neuropathy at the elbow using electrodiagnostic tests and nerve ultrasound.

The addition of ultrasound (US) to electrodiagnostic (EDX) tests can significantly enhance the accuracy of testing for ulnar neuropathy at the elbow (UNE). We aimed to obtain expert consensus to guide clinicians on the combined use of EDX and US in UNE investigation. Consensus was achieved using the Delphi method. Two consecutive anonymised questionnaires were submitted to 15 experts, who were asked to choose their level of agreement with each statement. Consensus was pre-defined as ≥ 80% rating agreement. The experts concluded that all investigations of UNE should include both nerve conduction studies and US. There was consensus that US should include cross-sectional area measurement and assessment of nerve mobility at the elbow, and that the entire ulnar nerve should be imaged. This study defined expert opinion on the 'core' techniques that should be used routinely in the UNE investigation using EDX and US. Areas with lack of consensus highlighted some controversial issues in the current use of these diagnostic modalities and the need for future research. This document is an initial step to guide clinicians on the combined investigation of UNE using EDX and US, to be regularly updated as new research emerges.

COVID-19 diagnosis by routine blood tests using machine learning.

Physicians taking care of patients with COVID-19 have described different changes in routine blood parameters. However, these changes hinder them from performing COVID-19 diagnoses. We constructed a machine learning model for COVID-19 diagnosis that was based and cross-validated on the routine blood tests of 5333 patients with various bacterial and viral infections, and 160 COVID-19-positive patients. We selected the operational ROC point at a sensitivity of 81.9% and a specificity of 97.9%. The cross-validated AUC was 0.97. The five most useful routine blood parameters for COVID-19 diagnosis according to the feature importance scoring of the XGBoost algorithm were: MCHC, eosinophil count, albumin, INR, and prothrombin activity percentage. t-SNE visualization showed that the blood parameters of the patients with a severe COVID-19 course are more like the parameters of a bacterial than a viral infection. The reported diagnostic accuracy is at least comparable and probably complementary to RT-PCR and chest CT studies. Patients with fever, cough, myalgia, and other symptoms can now have initial routine blood tests assessed by our diagnostic tool. All patients with a positive COVID-19 prediction would then undergo standard RT-PCR studies to confirm the diagnosis. We believe that our results represent a significant contribution to improvements in COVID-19 diagnosis.

Patterns and parameters describing nerve thickening in compression and entrapment ulnar neuropathies at the elbow.

OBJECTIVE: To compare pattern and parameters describing nerve thickening in ulnar neuropathy at the elbow (UNE) due to external compression in the retrocondylar groove (RTC), and entrapment under the humeroulnar aponeurosis (HUA). METHODS: In a group of our previously reported UNE patients we ultrasonographically (US) measured ulnar nerve cross-sectional areas (CSA) on 6-8 standard locations in the elbow segment. We compared CSA patterns in both groups, and determined diagnostic utility of selected CSA based parameters. RESULTS: We studied 79 patients (81 arms) with UNE due to external compression, and 53 patients (55 arms) due to entrapment. Maximal ulnar nerve CSA (>16 mm2), maximal CSA change (>7 mm2/1-2 cm) and maximal/minimal CSA ratio (>2.6) were significantly larger in UNE due to entrapment. They also differentiated these arms from arms with compression with sensitivities of 78%, 87% and 80%, and specificities of 90%, 94%, and 85%, respectively. CONCLUSION: Maximal difference in CSA between points separated by 1-2 cm (>7 mm2/1-2 cm) very efficiently differentiated between UNE due to external compression and entrapment. SIGNIFICANCE: The proposed parameter will hopefully complement precise localization in determining underlying mechanism of UNE. This may help physicians to determine the most appropriate treatment for UNE and possibly other focal neuropathies of unknown cause; i.e., conservative treatment for external compression and surgery for entrapment.

Differentiation of ulnar neuropathy at the wrist due to ganglion cyst from ulnar neuropathy at the elbow.

OBJECTIVE: Here, we aimed to describe the clinical, electrodiagnostic (EDx) and ultrasonographic (US) findings in a series of patients with ulnar neuropathy at the wrist (UNW) due to compression by a ganglion cyst. We also sought features that differentiate UNW from ulnar neuropathy at the elbow (UNE). METHODS: We reviewed electronic medical records of consecutive patients with UNW caused by ganglion cysts. We compared their clinical, EDx and US findings to findings in our previously reported prospective series of UNE patients. RESULTS: We identified 10 patients with UNW caused by ganglion cyst compression, who all presented with intrinsic hand muscle weakness and atrophy. Compared to 175 UNE patients they less often complained of paresthesia (60% vs. 98%) and presented less sensory loss in the palm (30% vs. 96%) and little finger (50% vs. 95%). They more often had distal ulnar motor latency recorded from the abductor digiti minimi (ADM)>3.6ms (80% vs. 30%), and denervation activity on needle EMG in the first dorsal interosseous (FDI) compared to ADM (100% vs. 60%). Only 20% of our UNW patients had ulnar nerve swelling at the site of compression on US. CONCLUSION: UNW potentially caused by ganglion cyst should be suspected in patients presenting with intrinsic hand muscle atrophy and weakness, particularly in cases with normal sensation, increased distal ulnar motor latency recorded from ADM and more severe neuropathic changes in FDI compared to ADM muscle.

Utility of nerve conduction studies and ultrasonography in ulnar neuropathies at the elbow of different severity.

OBJECTIVE: To report the sensitivity and the ability to precisely localize ulnar neuropathies at the elbow (UNE) of different severity by ultrasonography (US) and compare it to standard 10-cm nerve conduction studies (NCSs), and 2-cm short-segment NCSs (SSNCSs) across the elbow. METHODS: In a group of consecutive UNE patients, a prospective and blinded study was performed. The evaluation included clinical examination, electrodiagnostic (EDx) and US studies. We compared US and NCSs for sensitivity and the ability to precisely localize the UNE of different clinical severity. RESULTS: We studied 202 affected arms of 197 UNE patients. Clinically very mild UNE was diagnosed in seven, mild in 43, moderate in 99 and severe in 53 arms. The sensitivities of SSNCSs were 14%, 67%, 93% and 100%, of 10-cm NCSs, 29%, 44%, 80% and 96%, and of US 14%, 47%, 59% and 89%, respectively. Precise UNE localization was possible using SSNCSs in 29%, 56%, 78% and 85%, and using US in 29%, 44%, 70% and 98%, respectively. CONCLUSION: The present study demonstrated that NCSs are more sensitive than US for the diagnosis of UNE of all clinical grades of severity. US was more efficient in localizing clinically severe, and SSNCSs in localizing mild or moderate UNE. SIGNIFICANCE: We recommend SSNCSs as the first confirmatory test in UNE across all grades of severity.

Contribution of ultrasonography in evaluating traumatic lesions of the peripheral nerves.

OBJECTIVE: To assess the indications for and utility of diagnostic ultrasonography (US) in a series of consecutive patients with suspected traumatic peripheral nerve lesions (TPNL). METHODS: We retrospectively reviewed the electronic medical records of consecutive patients referred from February 2013 to May 2018 to our US laboratory. All included patients were examined using standard US equipment, with a 4-13MHz linear array transducer. RESULTS: In the analyzed period, we performed US examinations in 143 patients with 149 suspected TPNL. Electrodiagnostically (EDx) complete TPNL were found in 63 (45%), partial in 59 (42%), and only demyelination (i.e., neurapraxia) in four (3%) patients. TPNL could not be confirmed in 14 (10%) patients. Nerve discontinuity was not demonstrated by US in any of the patients with EDx incomplete nerve lesions. Contact of the nerve with osteosynthetic material (OSM) was found in eight of 33 patients (24%). In five patients, the nerve could not be adequately evaluated throughout its course due to extensive changes in the surrounding tissues. DISCUSSION: In acute situations, US is most useful in EDx complete TPNL to differentiate between nerve axonotmesis and neurotmesis. High-velocity trauma, lacerations, and bone fractures are all risk factors for neurotmesis. In chronic situations, US is useful in cases of functionally inefficient reinnervation, neuropathic pain, or progressive nerve dysfunction. In such patients, the surrounding tissues and the relation of the nerve to any OSM need to be carefully examined. US examination is probably not needed in patients with TPNL following acute blunt trauma, only minor clinical deficits and/or slightly/moderately abnormal EDx findings.

Diagnosing brain tumours by routine blood tests using machine learning.

Routine blood test results are assumed to contain much more information than is usually recognised even by the most experienced clinicians. Using routine blood tests from 15,176 neurological patients we built a machine learning predictive model for the diagnosis of brain tumours. We validated the model by retrospective analysis of 68 consecutive brain tumour and 215 control patients presenting to the neurological emergency service. Only patients with head imaging and routine blood test data were included in the validation sample. The sensitivity and specificity of the adapted tumour model in the validation group were 96% and 74%, respectively. Our data demonstrate the feasibility of brain tumour diagnosis from routine blood tests using machine learning. The reported diagnostic accuracy is comparable and possibly complementary to that of imaging studies. The presented machine learning approach opens a completely new avenue in the diagnosis of these grave neurological diseases and demonstrates the utility of valuable information obtained from routine blood tests.

Validation of clinical criteria for referral to head imaging in the neurologic emergency setting.

BACKGROUND: In recent decades, diagnostic imaging became an important generator of large increases in medical spending. Inappropriate head CT referrals also increase population irradiation and unnecessarily burden and frighten patients. OBJECTIVE: To validate previously proposed clinical criteria for referral to head imaging (age > 55 years, focal neurological deficit, changed mental state, nausea or vomiting, coagulation disorder, cancer) in a setting of emergency neurological service. METHODS: We retrospectively analyzed electronic records of 500 consecutive referrals to neurological emergency and 500 referrals to emergency head imaging. In patients with several referrals, only results of the first evaluation were further analyzed. We calculated relations between clinical predictors, referrals, and findings of head imaging. RESULTS: Of 486 first referrals of consecutive patients, 216 (44%) were referred to the emergency, and 100 (21%) to non-emergency head imaging. Remaining 170 (35%) were not referred to head imaging. Clinical predictors of pathologic head imaging fulfilled 77%, 41%, and 43% of patients, respectively. Pathologic head imaging had 153 of 490 (31%) referred patients. Referral criteria fulfilled 146 (sensitivity 95%) of them. Intracranial pathology was found in 7 of 125 patients not fulfilling referral criteria (negative predictive value 94%): 3 reported transient neurological symptoms, 2 sudden headache, and 2 headache with nausea and vomiting. CONCLUSION: We confirmed utility of previously proposed clinical criteria for referral to head CT in emergency neurological setting. In addition, we found transient neurological symptoms, sudden severe headache, and headache with nausea or vomiting as additional independent indications for emergency head imaging.

Standards for quantification of EMG and neurography.

This document is an update and extension of ICCN Standards published in 1999. It is the consensus of experts on the current status of EMG and Neurography methods. A panel of authors from different countries with different approach to routines in neurophysiological methods was chosen based on their particular interest and previous publications. Each member of the panel submitted a section on their particular area of interest and these submissions were circulated among the panel members for edits and comments. This process continued until a consensus was reached. The document covers EMG topics such as conventional EMG, Macro EMG, applications of surface EMG and electrical impedance myography. Single Fiber EMG is not included, since it is the topic in a separate IFCN document. A neurography section covers topics such as motor and sensory neurography, F wave recordings, H-reflex, short segment recordings, CMAP scan and motor unit number methods. Other sections cover repetitive nerve stimulation and Pediatric electrodiagnostic testing. Each method includes a description of methodologies, pitfalls, and the use of reference values. Clinical applications accompany some of these sections.

Long-term outcomes in patients with ulnar neuropathy at the elbow treated according to the presumed aetiology.

OBJECTIVES: Ulnar neuropathy at the elbow (UNE) consists mainly of two conditions: entrapment under the humeroulnar aponeurosis (HUA) and extrinsic compression in the retrocondylar (RTC) groove. These in our opinion need different treatment: surgical HUA release and avoidance of inappropriate arm positioning, respectively. We treated our UNE patients accordingly, and studied their long-term outcomes. METHODS: We invited our cohort of UNE patients to a follow-up examination consisting of history, neurological, electrodiagnostic (EDx) and ultrasonographic (US) examinations performed by four blinded investigators. RESULTS: At a mean follow-up time of 881 days, we performed a complete evaluation in 117 of 165 (65%) patients, with 96 (90%; 35 HUA and 61 RTC) treated according to our recommendations. An improvement was reported by 83% of HUA and 84% of RTC patients. In both groups the ulnar nerve mean compound muscle action potential (CMAP) amplitude, and the minimal motor nerve conduction velocity increased, while the maximal ulnar nerve cross-sectional area (CSA) decreased. CONCLUSION: After 2.5 years similar proportions of HUA and RTC patients reported clinical improvement that was supported by improvement in EDx and US findings. SIGNIFICANCE: These results suggest that patients with UNE improve following both surgical decompression and non-operative treatment. A clinical trial comparing treatment approaches in neuropathy localised to the HUA and RTC will be needed to possibly confirm our opinion that the therapeutic approach should be tailored according to the presumed aetiology of UNE.

Anorectal Dysfunction in Presymptomatic Mutation Carriers and Patients with Huntington's Disease.

BACKGROUND: Huntington's disease (HD) patients often report anorectal dysfunction; however, in HD research no detailed analysis of these complaints has been published. OBJECTIVE: To report anorectal dysfunction in a systematically studied cohort of HD subjects. METHODS: In 54 HD patients (24 men) and 10 presymptomatic HD mutation carriers (2 men) and in 99 controls (44 men) a history of anal incontinence and constipation was obtained and data was compared accordingly. In HD mutation carriers a clinical neurologic assessment and in some cases anorectal manometry were performed. RESULTS: Defecation urgency was reported by 28% of our HD mutation carriers, soiling in 18% and fecal incontinence in 28%. Severe anal incontinence (solid stools) was found in 0% men / 10% women, moderate (liquid stools) in 21% / 13%, and mild (flatus only) in 67% / 47% of our HD subjects. Compared to controls, anal incontinence was significantly more common in HD subjects (p < 0.001). Severe chronic constipation was found in 4.2% men / 0.0% women, moderate in 8.3% / 0.0%, and mild in 21% / 27% of HD subjects. Constipation was more common in HD men (p = 0.02) than in HD women (p = 0.144). Anorectal dysfunction was not reported by 54% of our HD subjects. Patients reporting incontinence or constipation were significantly more depressed (r = 0.53, p = 0.001). Upon anorectal manometry reduced resting anal pressure was found in 4 of 6 HD women. CONCLUSIONS: Our study demonstrated significant bowel dysfunction in HD patients. We propose these symptoms to be of central autonomic origin, although we cannot exclude effects of medication. These often neglected symptoms in HD subjects require greater attention from physicians.