Artificial intelligence-guided, single-lead EKG may be a game-changer for symptom-to-balloon time reduction in ST-elevated myocardial infarction

Background

Over decades, efforts to shave off life-saving minutes from ST-Elevated Myocardial Infarction (STEMI) care centred on reducing door-to-needle and door-to-balloon times. We firmly believe that symptom-to-balloon time should prove a better focus to this end. Challenges come with this goal as it heavily relies on a patient's perception and initiative to seek care, which we deem intelligent and wearable Artificial Intelligence (AI)-driven Single Lead EKG technologies as an attractive solution in modern-day cardiology.

Purpose

To provide an accurate, accessible, and cost-effective AI-driven Single Lead STEMI detection algorithm that can be embedded into wearable devices and employed in a self-administered fashion.

Methods

Database: EKG records from Mexico, Colombia, Argentina, and Brazil from April 2014 to December 2019. Dataset: A total of 11,567 12-lead EKG records of 10[s] length with a sampling frequency of 500 Hz, including the following balanced classes: angiographically confirmed and unconfirmed STEMI, branch blocks, non-specific ST-T abnormalities, normal and abnormal (200+ CPT codes, excluding those mentioned above). Cardiologists manually checked the label of each record to ensure precision. Pre-processing: We discard the first and last 250 samples as they may contain a standardisation pulse. The study applied a digital low pass filter of order 5 with a frequency cut-off of 35 Hz. The mean was subtracted from each Lead. Classification: The determined classes were “STEMI” (Including STEMI in different locations of the myocardium – anterior, inferior, and lateral); and “Not-STEMI” (Combination of randomly sample, branch blocks, non-specific ST-T changes, and abnormal records – 25% of each). Training and Testing: A 1-D Convolutional Neural Network was trained and tested with a dataset proportion of 90/10, respectively. A different model was trained and tested for each Lead, using the central 4,500 samples of the records. The last dense layer outputs a probability for each report of being STEMI or Not-STEMI. Lead V2 showed the best overall results. The model was further tested through the same methodology using the best Lead with a subset of the previous data, excluding the unconfirmed STEMI EKG records (Total 7,230 12-lead EKG records for Confirmed Only STEMI dataset). Performance metrics were reported for each experiment and compared.

Results

Combined STEMI data: Accuracy: 91.2%; Sensitivity: 89.6%; Specificity: 92.9%. Confirmed STEMI Only dataset: Accuracy: 92.4%; Sensitivity: 93.4%; Specificity: 91.4% (Figure 1).

Conclusion

By assiduously improving the quality of the model's input, we continue to assess our algorithm's performance and reliability for future clinical validation as a potential remote monitoring and early STEMI detection device.

Funding Acknowledgement

Type of funding sources: None.

Performance metrics of AI-enhanced single lead EKG maintained after entry of organised clustered data

Background

Our experience in creating innovative Artificial Intelligence-guided single lead EKG methodologies for ST-Elevation Myocardial Infarction (STEMI) detection within complex EKG records has been previously validated.

Purpose

By expanding the intricate variables of our previously tested algorithm input, we seek to further improve our STEMI detecting tool.

Methods

11,567 12-lead EKG records (10-s length, 500 Hz sample frequency) derived from the Latin America Telemedicine Infarct Network database from April 2014 to December 2019. From these records, we included the following balanced classes: angiographically confirmed and unconfirmed STEMI (divided by wall affected), branch blocks, non-specific ST-T changes, normal, and abnormal (Remaining 200+ CPT codes). Cardiologist annotations ensured precision (Ground truth). Determined classes were “STEMI” and “Not-STEMI”. A 1-D Convolutional Neural Network model was trained and tested for each lead with dataset proportions of 90/10, respectively. The last dense layer outputs a probability for each record being STEMI/Not-STEMI. The analysis also included performance metrics and false-negative reports.

Results

Overall, the most promising Single lead for STEMI detection was V2 (91.2% Accuracy, 89.6% Sensitivity, and 92.9% Specificity). 55% of false negatives were inferior wall STEMI (Table 1).

Conclusion

Appreciable progress of our new methodology compared to our previous experiences in AI-guided Single Lead for STEMI detection, especially for lead V2. By performing a thorough analysis of false-negative reports, we aspire to identify potential areas of STEMI detection weakness which will become the focus of future ventures.

Funding Acknowledgement

Type of funding sources: None.

No need for a cardiologist for AMI diagnosis – progress of transforming a behemoth telemedicine program with artificial intelligence

Background

The Latin American Telemedicine Infarct Network (LATIN) Telemedicine is a mammoth hub and spoke model that provides an umbrella of AMI protection for 100 million patients. In the program, 826,043 patients had a telemedicine encounter; 7,400 with AMI were diagnosed; 4,332 of them managed with guidelines-based strategies. We have gradually begun implementing a system for using Artificial Intelligence (AI) algorithms embedded into EKGs for rapid and accurate STEMI detection and validated the results with a cardiologist's interpretations.

Purpose

To test whether an AI-driven EKG algorithm can effectively substitute a cardiologist for STEMI telemedicine protocols.

Methods

The AI algorithm construction was in the following fashion. Sample: a selection of 8,511 EKG and 90,592 classified heartbeats. Pre-processing: segmentation of each EKG into individual heartbeats. Training & testing: 90% and 10% of the total dataset, respectively. Classification: 1-D Convolutional Neural Network; the study constructed classes for each heartbeat. The algorithm was next deployed on a consecutive series of LATIN EKG records to diagnose STEMI. We afterwards compared the algorithm's results with eight expert cardiologists' interpretations of the same sample.

Results

This study achieved a concordance of 91% between the AI algorithm and cardiologist interpretation (Figure 1).

Conclusions

The initial results with AI algorithms for STEMI diagnosis are encouraging and may provide the base work for new tools for cardiologists to improve their efficiency. Moreover, implementing this innovative tool may overcome current limitations associated with the telemedical management of this disease.

Funding Acknowledgement

Type of funding sources: None.

Abstract P4-07-02: Withdrawn

This abstract was withdrawn by the authors.

Citation Format: Katzenellenbogen BS, Guillen VS, Ziegler Y, Kim SH, Laws MJ, Zhao Y, Yasuda MA, Li Z, El-Ashry D, Katzenellenbogen JA. Withdrawn [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 P4-07-02.

An Internet-based self-help treatment for fear of public speaking: a controlled trial

This study offers data about the efficacy of "Talk to Me," an Internet-based telepsychology program for the treatment of fear of public speaking that includes the most active components in cognitive-behavior therapy (CBT) for social phobia (exposure and cognitive therapies). One hundred twenty-seven participants with social phobia were randomly assigned to three experimental conditions: (a) an Internet-based self-administered program; (b) the same program applied by a therapist; (c) a waiting-list control group. Results showed that both treatment conditions were equally efficacious. In addition, Talk to Me and the same treatment applied by a therapist were more efficacious than the waiting-list condition. Treatment gains were maintained at 1-year follow-up. The results from this study support the utility of Internet-delivered CBT programs in order to reach a higher number of people who could benefit from CBT. Internet-delivered CBT programs could also play a valuable role in the dissemination of CBT.

PDA self-register system for eating disorders: a study on acceptability and satisfaction

Self-monitoring techniques, such as the use of dietary registers, are considered to be central to cognitive-behavioral treatment of Eating Disorders (ED). This information allows the clinician to identify the triggers of the behaviors associated to ED as purges and/or binges, and the associated thoughts and emotions, helping to carry out a more accurate assessment. Traditionally these registers are made with paper and pencil mode, where the patient has to register every eating and the emotions/thoughts associated; but this system has some problems, as low portability, low adherence or methodological difficulties. The use of PDA for self-registers can help to solve these problems. The aim of this study is to study the levels of acceptability and satisfaction with PDAs self-register system specifically designed for assessment and treatment of ED. Samples of 30 subjects diagnosed with ED are receiving a PDA with software specifically designed for recording type and amount of food, emotions before and after eating and other behaviors. The participants are completing self-register daily during a week, and afterwards answer an acceptance and satisfaction questionnaire. This work is in progress at the moment. It is expected that the PDA system will show high levels of acceptance and satisfaction.

[Psychotherapy versus drug therapy in depression in ambulatory care]

Depression in Europe has a prevalence rate of 3.9%. One of the main work loads in out-patient care comes from treatment of affective disorders. The objective of the present study is to compare the efficacy of psychotherapy versus drug therapy in the treatment of affective disorders. The systematic review carried out has found 6 randomized controlled trials with a pill-placebo control group. The conclusions obtained after re-analyzing each study point out to comparatively equal efficacy of the active treatments and placebo in mild depressions. On the other hand, no significant differences were observed in relationship to the psychotherapeutic treatment efficacy versus drug treatments in moderate and severe depressions, these out-performing placebo efficacy.

Nodular regenerative hyperplasia of the liver and Hodgkin's disease: a case report

Nodular regenerative hyperplasia (NRH) of the liver is a rare entity characterized by the presence of nodules in the hepatic parenchyma, not surrounded by fibrous septa. The pathogenesis and etiology are unknown but an association with different diseases including some hematological disorders has been described. Its association with Hodgkin's disease is infrequent. We report the case of a 63 years old man who presented symptoms and signs of portal hypertension, hepatocellular failure with progressive deterioration and death. Postmortem examination disclosed Hodgkin's disease with hepatosplenic involvement and NRH of the liver. The association of these entities could be explained by the presence of portal infiltration contributing to portal venous obliteration and leading to portal hypertension and formation of the parenchymal nodules characteristic of this entity. Other mechanisms that could cause or influence this association can not be ruled out.

Absence of static phase separation in the high T(c) cuprate YBa(2)Cu(3)O(6+y)

We use 89Y NMR in YBa(2)Cu(3)O(6+y) in order to evaluate with high sensitivity the distribution of hole content p in the CuO2 planes. For y=1 and y=0.6, this hole doping distribution is found narrow with a full width at half maximum smaller than Deltap=0.025. This rules out any large static phase separation between underdoped and optimally doped regions in contrast with the one observed by STM in Bi2212 and by NQR in LaSrCuO. This establishes that static electronic phase separation is not a generic feature of the cuprates.