應用分割模型自動化量測水腦症影像指標

Abstract

水腦症為神經外科常見問題，由腦脊髓液過多且不正常堆積於腦部造成腦室異常擴大，引起顱內壓增高、頭痛、失禁、反應遲鈍等症狀。 而阻塞狀況分類可分為阻塞性及交通性水腦症，常見原因為先天性、感染、腫瘤或頭部外傷，臨床用於影像評估水腦症的參數有Evans ratio、Third Ventricle Width、Cella Media Index、Huckman’s Index和Bicaudate Index等,一旦確定診斷後，再依情況進行藥物或手術治療。然而，這些參數通常為臨床醫師於影像上手動量測，耗時且容易受個體差異的影響，我們將以分割模型標記腦組織，並結合演算法以自動化量測五種腦積水相關指標。

Hydrocephalus is a common neurosurgical condition characterized by the excessive and abnormal accumulation of cerebrospinal fluid (CSF) within the brain, leading to ventricular dilation.This dilation can result in symptoms such as increased intracranial pressure, headaches, incontinence, and delayed responses. Hydrocephalus can be classified as either obstructive or communicating,with common causes including congenital factors, infections, tumors, or head trauma. Clinically, imaging evaluation of hydrocephalus often involves parameters like the Evans Ratio, Third Ventricle Width, Cella Media Index, Huckman's Index, and Bicaudate Index.Once a diagnosis is confirmed, treatment typically involves medication or surgery. However,these parameters are traditionally measured manually by clinicians on medical images.This process is not only time-consuming but also susceptible to inter-individual variability.Our approach aims to address these limitations by utilizing segmentation models to delineate brain tissues and integrate algorithms for the automated measurement of these five hydrocephalus-related indicators.