P1464Adoption of feedback to validate a machine learning model for single lead STEMI detection

Background

We have explored the performance of a single lead EKG with Artificial Intelligence (AI) based algorithms in STEMI diagnosis, thus far lead V2 has yielded the best results. Anticipating the performance of the LUMENGT-AI model, we designed a feedback strategy with healthcare centers to expand the validation of our work.

Purpose

To create a pragmatic alternative to the existing gold standard, a 12-lead EKG, for STEMI diagnosis.

Methods

An observational, retrospective, case-control study. Sample: 2,543 exclusively STEMI (anterior, inferior and lateral wall) diagnosis, EKG records. Feedback: From healthcare centers, confirming STEMI diagnosis and location, was obtained (thrombolysis, primary Percutaneous Coronary Intervention (PCI), pharmaco invasive therapy or coronary artery bypass surgery). Records excluded other patient and medical information. Sample was derived from the private International Telemedical Systems (ITMS) database. LUMENGT-AI Algorithm was employed. Preprocessing: detection of QRS complexes using the wavelet system, segmentation of each EKG into individual heartbeats (53,667 total beats) with fixed window of 0.4s to the left and 0.9s to the right of main QRS; Classification: A 1-D convolutional neural network was implemented; “STEMI” and “Not-STEMI” classes were considered for each heartbeat per lead; individual probabilities were aggregated to generate the final label for each record. Training & Testing: 90% and 10% of the sample were used, respectively. Experiments: Intel PC i7 8750H processor at 2.21GHz, 16GB RAM, Windows 10 OS with NVidia GTX 1070 GPU, 8GB RAM.

Results

V2 was the most precise lead with an Accuracy of 93.6%, a Sensitivity of 89%, and a Specificity of 94.7%.

Conclusions

The strategic adoption of feedback from healthcare centers provided strong validation of our model. The results of AI-augmented, single lead EKG are encouraging. We anticipate that this approach will become a promising methodology in STEMI detection.

P1466Can physicians trust a machine learning algorithm to diagnose ST elevation myocardial infarction?

Background

For the past years, the medical field has been taking advantage of the endless possibilities that Artificial Intelligence (AI) provides. Using computer-aided devices that can perform and interpret electrocardiograms (EKG) accurately pushes current healthcare boundaries. We present the LUMENGT-AI, this model can handle large datasets, multiclass diagnoses, complex EKG morphology, and still detect ST Elevation MI (STEMI) accurately.

Purpose

To develop an innovative AI-based system for automated STEMI specific EKG analysis.

Methods

An observational, retrospective, case-control study. Sample: 8,511 EKG records, previously diagnosed as “normal”, “abnormal” (over 200 conditions) or “STEMI” (4,255 cases). Records excluded patient and medical information. The sample was derived from the private International Telemedical Systems (ITMS) database. LUMENGT-AI Algorithm was employed. Preprocessing: detection of QRS complexes by wavelet system, segmentation of each EKG into individual heartbeats (90,592 total beats) with fixed window of 0.4s to the left and 0.9s to the right of main QRS; Classification: A 1-D convolutional neural network was implemented, “STEMI” and “Not-STEMI” classes were considered for each heartbeat, individual probabilities were aggregated to generate the final label for each record. Training & Testing: 90% and 10% of the sample were used, respectively. Experiments: Intel PC i7 8750H processor at 2.21GHz, 16GB RAM, Windows 10 OS with NVidia GTX 1070 GPU, 8GB RAM.

Results

Ground Truth Score – Accuracy (94.1%), Sensitivity (87.8%), Specificity (98.1%) – see the comparison to published data in Table.

Conclusions

A statistical analysis allowed us to compare STEMI recognition efficiency between physicians and our model. The LUMENGT algorithm results secured its place as a reliable tool to diagnose STEMI faster and more accurately than physicians.

Exposure-safety and efficacy response relationships and population pharmacokinetics of eslicarbazepine acetate

Objectives

Eslicarbazepine acetate (ESL) is a once‐daily (QD) oral antiepileptic drug (AED) for focal‐onset seizures (FOS). Pharmacokinetic (PK) and pharmacodynamic (PD) models were developed to assess dose selection, identify significant AED drug interactions, and quantitate relationships between exposure and safety and efficacy outcomes from Phase 3 trials of adjunctive ESL.

Methods

Eslicarbazepine (the primary active metabolite of ESL) population PK was evaluated using data from 1351 subjects enrolled in 14 studies (11 Phase 1 and three Phase 3 studies) after multiple oral doses ranging from 400 to 1200 mg. Population PK and PD models related individual eslicarbazepine exposures to safety outcomes and efficacy responses.

Results

Eslicarbazepine PK was described by a one‐compartment model with linear absorption and elimination. The probability of a treatment‐emergent adverse event (TEAE; dizziness, headache, or somnolence) was higher with an initial dose of ESL 800 mg than with an initial dose of ESL 400 mg QD. Body weight, sex, region, and baseline use of carbamazepine (CBZ) or lamotrigine were also found to influence the probability of TEAEs. Eslicarbazepine exposure influenced serum sodium concentration, standardized seizure frequency, and probability of response; better efficacy outcomes were predicted in patients not from Western Europe (WE; vs WE patients) and those not taking CBZ (vs taking CBZ) at baseline.

Conclusions

Pharmacokinetic and PK/PD modeling were implemented during the development of ESL for adjunctive treatment of FOS in adults. This quantitative approach supported decision‐making during the development of ESL, and contributed to dosing recommendations and labeling information related to drug interactions.

Chronic rhinosinusitis and mood disturbance

BACKGROUND

This study is part of the Chronic Rhinosinusitis Epidemiology Study (CRES). The overarching aim is to determine factors that influence the onset and severity of chronic rhinosinusitis (CRS). The aim of this analysis is to determine whether those with CRS are more likely to report psychiatric morbidity and in particular mood disturbance compared with healthy controls.

METHODS

CRES consists of a study-specific questionnaire regarding demographic and socioeconomic factors and past medical history as well as a nasal symptom score (SNOT-22) and SF-36 (QoL - quality of life tool). Both of these tools contain mental health or emotional well-being domains. Participants were specifically asked whether they had ever consulted with their General Practitioner for anxiety or depression. Questionnaires were distributed to patients with CRS attending ENT outpatient clinics at 30 centres across the United Kingdom from 2007-2013. Controls were also recruited at these sites. Patients were divided into subgroups of CRS according to the absence/presence of polyps (CRSsNPs/CRSwNPs) or allergic fungal rhinosinusitis (AFRS).

RESULTS

Consultations with a family physician for depression or anxiety were higher amongst those with CRS than controls, but this was only significant for those with CRSsNPs. Odds ratio (OR) for CRSsNPs vs controls: 1.89; OR for CRSwNPs: 1.40. Patients with CRS showed significantly higher mental health morbidity than controls across the mental health and emotional wellbeing domains of the SF-36 and SNOT-22. Mean difference in the mental health domain of SF-36 was 8.3 for CRSsNPs and 5.3 for CRSwNPs. For the emotional domain of SNOT-22, differences were 7.7 and 6.3 respectively.

CONCLUSIONS

Depression and anxiety are significantly more common in patients with CRS compared to healthy controls, especially in those with CRSsNPs. This added mental health morbidity needs consideration when managing these patients in primary and secondary care settings.