Skripsi

PENILAIAN VIABILITAS FREE FLAP BERDASARKAN CITRA 2-D PENDEKATAN ARTIFICAL INTELLIGENCE DENGAN VALIDASI KONDISI KLINIS, NILAI LAKTAT DAN GLUKOSA

---

Background: Reconstruction of defect closure following tumor ablation using free flaps represents a superior option; however, this procedure is technically challenging and carries a high risk of failure. Clinical examination remains the gold standard for assessing free flap viability, yet it is inherently subjective and dependent on the examiner’s experience. Therefore, an objective and universally applicable assessment tool is required. Biomarkers such as lactate and glucose levels can serve to detect ischemia. 2-D imaging with Artificial Intelligence (AI) approaches holds promise as an innovative solution to consistently evaluate free flap viability, validated against clinical findings as well as lactate and glucose measurements. Methods: This study employed a diagnostic test design and concordance testing with a cross-sectional approach. The research subjects comprised 32 free flap patients at Dr. Mohammad Hoesin Hospital, Palembang. The study was conducted in four stages: (1) clinical examination and acquisition of 2-D images; (2) measurement of lactate and glucose levels at five time points; (3) development of the Artificial Intelligence (AI) model; and (4) validation of the AI model against the assessments of three Microvascular and Oncoplastic Surgery Consultants. Results: Thirty-two subjects produced 5,804 two-dimensional (2-D) images. Lactate values were not significant in differentiating viable from compromised free flaps, whereas glucose values predicted compromised flaps early, with levels ≤ 76.5 mg/dL indicating compromised viability. This finding established the biomedical pathway for free flap assessment. The Vision Transformer (ViT) AI model with BS10-LR4 parameters achieved 100% accuracy, precision, sensitivity, and specificity. Validation against three Microsurgeon showed accuracy of 94%, precision 92%, sensitivity 62%, and specificity 99%, indicating comparable performance. Multivariate analysis demonstrated that lactate and glucose both contributed to visual changes recognized by the ViT AI model. Thus, the ViT AI model detects visual patterns while reflecting consistent metabolic dynamics captured by these biomarkers. Conclusion: Glucose values can serve as an objective indicator of free flap viability. This study also successfully mapped the biomedical pathway of free flaps. Two-dimensional (2-D) imaging using the Vision Transformer (ViT) AI model was able to assess free flap viability, reflecting both visual patterns and metabolic dynamics derived from clinical conditions, lactate, and glucose levels. Furthermore, this approach has the potential to be developed into FID (Flap in Detection) software, enabling direct application in patient care.

---

Availability

Detail Information

Series Title

-

Call Number

T1900142025

Publisher

Indralaya : Prodi Magister Ilmu Biomedik, Fakultas Kedokteran Universitas Sriwijaya., 2026

Collation

xiii, 80 hlm.; ilus.; tab.; 29 cm.

Language

Indonesia

ISBN/ISSN

-

Classification

617.952 07

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Prodi Magister Ilmu Biomedik
Bedah Rekonstruksi

Specific Detail Info

-

Statement of Responsibility

TUTI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment