|
|
Am. J. Biomed. Sci. 2022,14(2),72-90;doi:10.5099/aj220200072 |
Towards Predicting
Polycystic Ovary Syndrome with a Novel Smartphone-based Biomedical Application Lyfas |
Subhagata Chattopadhyay*1, Rupam Das1 |
1 Department of Research and
Development (Digital Health), Acculi Labs Pvt. Ltd.
Bangalore 560098, Karnataka India |
*Corresponding
Author |
Subhagata Chattopadhyay |
Department of Research and Development (Digital
Health) Acculi Labs Pvt. Ltd. Bangalore 560098,
Karnataka |
India |
Email: subhagata.chattopadhyay2017@gmail.com |
Phone Number: +919972774547
|
Abstract Background: Polycystic Ovary Syndrome (PCOS) adversely affects reproductive and metabolic health. It mandates early detection. Lyfas is a mHealth instrument, which is a personalized, fast, non-invasive, and pervasive smartphone-based application. It captures Heart rate variability and its biomarkers (HRVBs) by finger touch on the phone camera. HRVBs are a surrogate for cardiac autonomic modulation that occurs in PCOS. |
Objective: Early prediction of PCOS by Lyfas HRVBs and its validation by gynecologists. |
Methods and Material: A retrospective double-blind control trial has been conducted on a mixed population of PCOS (N=218) and healthy (N=153) participants. The cohort is further divided into a) ‘Forward miners or FM’ (N=210: PCOS 135, healthy 75), where Lyfas has been used to mine the ‘significant’ HRVBs of PCOS, and b) ‘Reverse mappers or RM’ (N=161), where Lyfas decisions, based on the ‘significant’ HRVBs are validated by a panel of gynecologists. |
Statistical analysis: Cronbach’s alpha, Descriptive statistics, Q-Q plots, Spearman’s correlation, and classification metric (recall, specificity, precision, accuracy, fscore, and Youden’s index), and Bland-Altman’s reliability test (BART). |
Results: LF/HF and SD1/SD2 shows significant positive correlation (ρ = 0.60 and 0.45 and p-value = 0.009 and 0.02, respectively). Lyfas shows 82% recall, 84% specificity, 85% precision, 83% accuracy, 84% fscore, and 74% Youden’s index when compared to the diagnoses of gynecologists. BART shows Lyfas has a 2% of proportional bias, i.e., 98% reliable when compared to gynecologists’ prediction. |
Conclusions: Lyfas HRVBs (LF/HF and SD1/SD2) can be assistive to gynecologists to predict the possibility of PCOS in the suspected population in 3-5 minutes. |
Keywords: Lyfas; Polycystic Ovary Syndrome; Cardiovascular optical biomarkers; Heart rate variability |
Download the full article (PDF)
|
Publisher | Missions and Scope | Editorial Board | Instructions for Authors |
© American Journal of Biomedical Sciences 2007-2021. All Rights Reserved. |