Pengembangan Sistem Deteksi Pemanjangan Interval Qt Untuk Pasien Mdr-tb Dengan Metode Adaptif Threshold

Rayhani, Kireina Nazwa (2026) Pengembangan Sistem Deteksi Pemanjangan Interval Qt Untuk Pasien Mdr-tb Dengan Metode Adaptif Threshold. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5049221012-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Pengobatan MDR-TB berisiko tinggi menyebabkan pemanjangan interval QT yang dapat memicu aritmia jantung fatal. Pemantauan EKG konvensional seringkali tidak mampu mendeteksi perubahan interval QT yang bersifat transien. Penelitian ini mengembangkan sistem deteksi otomatis berbasis kombinasi algoritma Pan-Tompkins yang dimodifikasi dan metode Cartesian Geometric Curvature (CGC) untuk penentuan Q-onset dan T-end secara detak ke detak, dengan koreksi QTcF menggunakan formula Fridericia. Sistem diuji pada rekaman EKG kanal V6 dari 45 pasien MDR-TB dan 43 kontrol sehat. Pada kelompok kontrol sehat, koreksi Fridericia terbukti beroperasi ideal dengan menekan eksponen sensitivitas QT-RR hingga mendekati nol (−0,08), sementara pada pasien MDR-TB eksponen awal yang jauh lebih curam (0,57) menyebabkan formula tersebut mengalami under-correction, menyisakan kemiringan residual sebesar 0,24. Meski demikian, rata-rata QTcF pasien MDR-TB (462±32 ms) lebih panjang secara signifikan dibandingkan kontrol sehat (441±21 ms) dengan uji Mann-Whitney U (p<0,001, effect size r=0,43), disertai pendataran gelombang T (T-wave flattening) sebesar 24,84%. Hasil ini membuktikan sistem mampu mendeteksi pemanjangan interval QT secara akurat dan adaptif terhadap variasi morfologi gelombang T. Keandalan sistem juga divalidasi terhadap PTB Diagnostic ECG Database sebagai standar referensi publik: uji Fisher r-to-z menunjukkan tidak terdapat perbedaan performa yang signifikan antara data primer dan database sekunder, baik pada kelompok kontrol (r=0,946 primer vs r=0,948 PTB Database, p=0,911) maupun kelompok pasien (r=0,656 primer vs r=0,644 PTB Database, p=0,904), mengonfirmasi kesetaraan statistik data primer terhadap standar referensi publik.
===============================================================================================================================
MDR-TB treatment carries a high risk of QT interval prolongation, potentially triggering fatal cardiac arrhythmias. Conventional ECG monitoring often fails to detect transient QT changes. This study develops an automated detection system combining a modified Pan-Tompkins algorithm and the Cartesian Geometric Curvature (CGC) method for beat-to-beat Q-onset and T-wave endpoint determination, with QTcF correction using the Fridericia formula. The system was tested on V6-lead ECG recordings from 45 MDR-TB patients and 43 healthy controls. In the healthy control group, Fridericia correction operated ideally, flattening the QT-RR sensitivity exponent to near zero (−0.08), whereas in MDR-TB patients a markedly steeper baseline exponent (0.57) caused the formula to under-correct, leaving a residual slope of 0.24. Nevertheless, mean QTcF in MDR-TB patients (462±32 ms) was significantly longer than in healthy controls (441±21 ms), confirmed by a Mann-Whitney U test (p<0.001, effect size r=0.43), accompanied by 24.84% T-wave flattening. These findings demonstrate that the system accurately detects QT prolongation and is highly adaptive to T-wave morphological variations. The system's reliability was further validated against the PTB Diagnostic ECG Database as a public reference standard: Fisher r-to-z tests showed no significant difference in model performance between primary and secondary data sources, both in the control group (r=0.946 primary vs. r=0.948 PTB Database, p=0.911) and the patient group (r=0.659 primary vs. r=0.644 PTB Database, p=0.904), confirming the statistical equivalence of primary data relative to the public reference standard.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Automated Biosignal Processing, Cartesian Geometric Curvature, Electrocardiography (ECG), MDR-TB, QT Interval Prolongation
Subjects:	R Medicine > RC Internal medicine R Medicine > RC Internal medicine > RC683.5.E5 Electrocardiography T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5102.9 Signal processing.
Divisions:	Faculty of medicine and health (MEDICS) > Medical Technology > 11503-(S1) Undergraduate Thesis
Depositing User:	Kireina Nazwa Rayhani
Date Deposited:	29 Jul 2026 01:11
Last Modified:	29 Jul 2026 01:11
URI:	http://repository.its.ac.id/id/eprint/139266

Actions (login required)

View Item