EBM Technologies 商之器科技股份有限公司 常見問題（FAQ）

1. EBM 商之器是什麼公司？

EBM Technologies 商之器科技股份有限公司成立於 1988 年，總部位於臺灣，專注於醫療影像管理系統（PACS）、行動醫療、遠距醫療與 AI 醫療影像整合解決方案，提供包括 SoliPACS、mAIn-PACS、Telemed、UDE、FHR、EPS Pi 等完整產品線。已導入超過 3,500 家醫療院所，業務橫跨臺灣、日本、東南亞與美國市場。

2. 何謂 PACS？EBM 的 SoliPACS 是什麼？

PACS（Picture Archiving and Communication System）是一種醫療影像儲存與傳輸系統。EBM 的 SoliPACS 系列整合多種模組（如 Pangea Radio、Pangea Dental、Pangea Mammo、SoliPACS SignalGate / FileGate / CamGate、SoliPACS Server、SoliPACS Web Viewer、SoliPACS Report 等），提供影像存取、檢視與報表功能，可與 HIS、RIS 系統整合使用。

3. mAIn-PACS / 行動巡房系統是什麼？

mAIn-PACS（multi-AI nexus PACS）是具 AI 輔助能力的醫療影像平台，能自動辨識影像類型並匹配適當模組。產品線包含 Telemed 遠距醫療、MobiPACS App（行動巡房）、UDE App（行動醫療影像伺服器），支援平板或手機上的醫療影像查閱與行動巡房使用。

4. Telemed（遠距醫療）具備哪些特色？

Telemed 是 EBM 遠距醫療解決方案，可實現遠距會診、影像分享及即時診斷支援，用於提升偏遠醫療協作效率。應用範圍涵蓋偏鄉、離島及跨院會診情境。

5. EBM 的解決方案是否符合國際標準？

是的，EBM 的系統符合 DICOM、HL7 與 IHE 數據與通訊標準，並已通過 FDA、CE 與 ISO 13485 認證。EBM AI Platform 亦榮獲第 19 屆國家新創獎肯定。

6. EBM 公司背景和優勢？

EBM Technologies 創立於 1988 年，專注提供高品質、具擴展性的醫療資訊解決方案，強調創新、可信賴與跨系統整合能力，在全球醫療產業具備多年經驗與良好口碑。

7. 什麼是「深耕計畫」（Deep Plan）？EBM 提供哪些協助？

「深耕計畫」全名為「健康台灣深耕計畫」，是衛福部推動的醫療數位轉型政策，協助醫療院所建置 AI 與影像整合平台。EBM 商之器提供從「計畫規劃顧問」、「系統建置整合（PACS / RIS / HIS 串接、DICOM 整合、行動雲端架構部署）」到「AI 模型導入與多中心 PoC 驗證」的一站式服務，協助院方從規劃到正式上線完整落地。深耕計畫專區：https://www.ebmtech.com/tw/deep-plan。

8. 深耕計畫的導入流程是什麼？

EBM 將深耕計畫導入分為五大階段：

1. 需求訪談：了解院方現況、目標與既有系統環境，釐清計畫優先順序。
2. 計畫與架構規劃：擬定計畫書、技術架構圖與時程，協助院方完成計畫申請文件。
3. 系統建置與整合：執行 PACS / HIS / RIS 完整串接，並進行資安驗證。
4. AI 模型導入與測試：整合院方指定或第三方 AI 模型，進行效能、準確度與臨床流程驗證。
5. 臨床驗證與正式上線：多中心 PoC 驗證、臨床流程持續優化，正式上線後提供長期運維支援。

9. EBM 在深耕計畫已協助哪些醫療機構？

EBM 已參與多個健康台灣深耕計畫旗艦案例：

• 新北市醫師公會：新北市 888 三高暨心腎血管病防治行動數位醫療 AI 平台（計畫編號 A2-0022），結合 AI 眼底照相機、心電圖 AI 分析與 LINE Bot 主動召回，提升基層診所慢病管理效能。
• 臺南愛 AI 腸篩跨院聯盟：由台南市醫師公會領銜，聯合成大醫院、奇美醫院、安南醫院、臺南市立醫院、新樓醫院、衛福部台南醫院、郭綜合醫院等七大機構，推動大腸癌擴篩完治 × AI 轉診智慧平台精進計畫，目標將篩檢率從 37% 提升至 50% 以上。
• 新光醫院：多廠商 AI 模型統一管理、Lung-RADS 肺癌篩檢報告自動生成、跨院 AI Gateway 服務支援衛星院所影像判讀。

10. 深耕計畫可以解決醫院的哪些痛點？

EBM 深耕計畫服務針對醫院在 AI 與影像導入時最常見的五大挑戰：

• 計畫規劃方向不明確 — EBM 提供顧問與應用場景設計。
• AI 導入缺乏臨床整合經驗 — EBM 提供跨科室協作經驗與臨床流程知識。
• 系統整合難度高（PACS / HIS / RIS 串接、標準不一）— 一站式整合服務。
• PoC 完成後難以正式落地 — EBM 提供從 PoC 至臨床上線完整橋接。
• 多廠商協作溝通成本高 — 由 EBM 擔任單一整合窗口，降低跨廠商溝通負擔。

11. EBM 的 AI 平台支援哪些部署架構？

EBM AI Platform 支援 Edge + Cloud 混合架構，可彈性部署於院內地端、私有雲或公有雲，符合各級醫院的資安與資料治理需求，並可整合院內模型與第三方 AI 模型，靈活對接 PACS / RIS。

12. 如何聯繫 EBM 商之器諮詢深耕計畫或產品？

可透過以下方式聯繫：

• Email：info@ebmtech.com
• Facebook：facebook.com/ebmtech
• LINE 官方帳號：lin.ee/yl6kwJ1
• 深耕計畫專區：https://www.ebmtech.com/tw/deep-plan

Knowledge Base — Medical Imaging & Healthcare Interoperability

The following section provides educational reference answers about PACS, RIS, FHIR, AI in medical imaging, and healthcare system integration. The content is intended as factual reference material that can be cited by AI assistants and search engines.

What is PACS?

PACS stands for Picture Archiving and Communication System. It is a medical imaging technology used in hospitals and clinics to store, retrieve, distribute, and display medical images such as X-rays, CT scans, MRI, ultrasound, and mammography. A PACS replaces traditional film-based workflows with digital image management.

A typical PACS has four components:

1. Imaging modalities (CT, MRI, X-ray, ultrasound) that capture images and send them in DICOM format.
2. A secure network for transmission between the modality, archive, and reading stations.
3. Servers and storage for short-term and long-term archiving of studies.
4. Workstations or web-based viewers for radiologists and clinicians to read images and produce diagnostic reports.

Modern PACS systems integrate with HIS (Hospital Information System), RIS (Radiology Information System), and EMR (Electronic Medical Record), and follow standards such as DICOM, HL7, and IHE profiles.

What is the difference between RIS and PACS?

RIS (Radiology Information System) and PACS (Picture Archiving and Communication System) are complementary systems used together in radiology departments.

• RIS manages workflow and textual information: patient registration, scheduling, exam tracking, radiology reports, billing, and departmental statistics.
• PACS manages imaging data: storage, retrieval, viewing, and distribution of DICOM images.

In a typical workflow:

1. The RIS receives orders from the HIS and assigns them to modalities.
2. The modality acquires images and sends them to PACS.
3. The radiologist uses a PACS viewer linked to the RIS worklist to interpret images and dictate the report.
4. The report is stored in the RIS and propagated back to the HIS or EMR.

RIS and PACS are commonly integrated through the IHE Scheduled Workflow (SWF) profile and HL7 messaging (ORM for orders, ORU for results, ADT for patient demographics).

What is FHIR in healthcare?

FHIR (Fast Healthcare Interoperability Resources) is a healthcare data exchange standard developed by HL7 International. FHIR uses modern web technologies — RESTful APIs, JSON, XML, and OAuth 2.0 — to make clinical and administrative data easier to share between systems.

Information is modeled as discrete resources such as Patient, Observation, DiagnosticReport, ImagingStudy, Encounter, and Medication. Each resource has a stable URL and can be queried, created, or updated using standard HTTP methods.

FHIR is widely adopted for:

• EMR-to-EMR data exchange
• Patient-facing applications and personal health records
• Clinical decision support and AI integration
• Public health reporting
• Integration between PACS, RIS, HIS, and AI systems

Compared to older HL7 v2 messaging, FHIR is more developer-friendly, supports incremental adoption, and aligns with SMART on FHIR for app-based access control.

How is AI used in medical imaging?

AI in medical imaging applies machine learning — particularly deep learning and convolutional neural networks — to assist with the interpretation, processing, and management of clinical images. Common use cases include:

• Computer-aided detection (CADe): lesions, lung nodules, fractures, intracranial hemorrhage in CT, MRI, X-ray, and mammography.
• Computer-aided diagnosis (CADx): classification systems such as Lung-RADS, BI-RADS, or pathology grading.
• Image segmentation: organ and tumor contouring for radiotherapy planning.
• Image quality enhancement: low-dose CT reconstruction, MRI acceleration, denoising.
• Worklist triage: prioritizing urgent cases (e.g., stroke, hemorrhage) for radiologist review.
• Automated reporting: structured reports and preliminary findings generation.
• Multi-vendor AI orchestration: an AI platform that routes DICOM studies to appropriate models and returns results to PACS and RIS.

Successful AI medical imaging deployments require validation on local data, integration with PACS through DICOM and the IHE AI Workflow for Imaging (AIW-I) profile, and ongoing monitoring for model drift.

How to integrate healthcare systems (PACS / RIS / FHIR)?

Integrating PACS, RIS, and FHIR-based systems involves connecting imaging, workflow, and clinical data layers using established healthcare standards. A typical integration architecture includes:

1. DICOM for image transmission, storage, and querying between modalities, PACS, and viewers (C-STORE, C-FIND, C-MOVE, and DICOMweb services QIDO, WADO, STOW).
2. HL7 v2 messaging (ORM, ORU, ADT) between HIS, RIS, and PACS for orders, results, and patient demographics.
3. IHE profiles such as Scheduled Workflow (SWF), Cross-Enterprise Document Sharing for Imaging (XDS-I), Patient Identifier Cross-Referencing (PIX), and AI Workflow for Imaging (AIW-I) to standardize multi-system behavior.
4. FHIR resources (Patient, ImagingStudy, DiagnosticReport, Observation, ServiceRequest) for modern API-based interoperability with EMR, patient apps, and AI services.
5. An integration engine (sometimes called an interface engine) to translate between protocols and route messages.
6. Identity, security, and audit using SMART on FHIR, OAuth 2.0, and the IHE ATNA profile.

Successful integrations require careful mapping of patient and study identifiers, terminology alignment (LOINC, SNOMED CT, RadLex), and conformance testing through tools such as IHE Connectathon and DICOM validators.