肌少症不只影響老年人的行動能力、生活品質，同時還會增加罹病率以及死亡率。現有診斷肌少症的方式，經常伴隨著高成本、儀器低可及性以及可能有輻射曝露等問題，且通常需到醫院接受一系列檢查，因此能促進方便性、即時性且具有準確性的肌少症診斷工具便是迫切且尚未滿足的醫療需求。
現有研究已發現循環系統疾病與肌少症之間存在關聯性。動脈血管特性變化會影響肌少症疾病的發展。本研究會藉由本實驗室自行開發的穿戴式脈波量測裝置來量測動脈脈波，藉以探討患有肌少症早期症狀族群(Possible sarcopenia，簡稱PSarco)與一般壯健民眾(Robust)於血管特性的差異，而為了達到更早期偵測肌少症的目的，亦會將肌少症的潛在因子 — 潛在代謝症候群(PreMetS)也納入與肌少症合併分析。
有133名前來老年健康檢查的民眾接受了雙盲實驗，進行了3分鐘非侵入式的血壓波形量測。經收案統計，可得到Robust有 74人；PSarco 有59人。本研究會採用脈波頻域分析和機器學習(ML)分析(以40個頻域諧波參數為特徵：能量比例(Cn)及其變異係數(CVn)，相位角(Pn)及其標準偏差(Pn_SD))來進行數據分析，並自行開發一個脈波評分系統 — 脈波分佈區間分析法 ( PDA )，來提供更輕便且快速的分析方式。
本研究結果發現，PSarco與Robust在眾多頻域諧波參數上具有顯著差異，尤其Cn在全部10個諧波皆具有顯著性。將PreMetS也納入合併分析，可發現伴隨著兩者症狀的發生，對於血管特性會有最嚴重的影響。基於這些基礎，在ML的表現，經三倍交叉驗證(3-fold cross-validation)結果，使用LDA分類器模型分類的性能AUC可高達0.74。而使用PDA的結果AUC最終可高達0.83，兩者皆已達到可實務應用的標準。
基於藉由觀察因肌少症引起的血管特性改變而造成動脈脈波傳輸條件變化的測量和分析，最終得到以下幾點結論：
1. 患有肌少症症狀的病患(PSarco)與一般壯健民眾(Robust)在眾多頻域諧波上具有顯著的差異性。
2. 與Robust相比，同時患有代謝症候群與肌少症症狀的受試者(PSarco_PreMetS)存在著最明顯的差異。
3. 使用ML分析，經3-fold cross-validation的結果，使用 LDA 分析時分類性能AUC可高達0.74，運算相較PDA更仔細且全面，隨著樣本數持續增加預計會有更好的表現，適合雲端運算。
4. 使用PDA分析，分類性能AUC可高達0.83，且運算簡單快速，適合裝置端運算。
5. 目前的量測與分析（包括 ML 分析和PDA分析）檢測肌少症引起的動脈脈波傳輸條件變化，有助於提供一種非侵入式且易於使用的方法，來判別可能肌少症(PSarco)，未來可望結合雲端架構，來提供使用者更全面性的服務。

Sarcopenia not only affects the mobility and quality of life of the elderly but also increases morbidity and mortality. Existing modalities for diagnosing sarcopenia are often associated with high cost, low instrument accessibility, and possible radiation exposure, and often require a series of tests in a hospital, thus facilitating convenient, timely, and accurate diagnosis. A diagnostic tool for sarcopenia is an urgent and unmet medical need.
Existing research has found an association between circulatory disorders and sarcopenia, with changes in arterial vascular properties affecting the development of sarcopenia. This study will use the wearable pulse wave measurement device developed in our laboratory to measure the arterial pulse wave, to explore the blood vessels of the early symptom group of possible sarcopenia (PSarco) and the general healthy people (robust). For early detection of sarcopenia, underlying metabolic syndrome (PreMetS), a potential sarcopenic factor, was also included in the combined analysis with sarcopenia.
A 3-minute non-invasive blood pressure waveform measurement was performed on 133 people who came to the elderly health check-up in a double-blind experiment. According to the statistics, there are 74 people in Robust and 59 people in PSarco. This research will use pulse wave frequency domain analysis and machine learning (ML) analysis (characterized by 40 frequency domain harmonic parameters: energy ratio (Cn) and its coefficient of variation (CVn), phase angle (Pn), and its standard deviation (Pn_SD)) for data analysis, and developed a pulse wave scoring system—Pulse Distribution Analysis (PDA) to provide a lighter and faster analysis method.
The results of this study found that PSarco and Robust have significant differences in many frequency-domain harmonic parameters, especially Cn is significant in all 10 harmonics. When PreMetS was included in the pooled analysis, it was found that the most severe impact on vascular properties was associated with the occurrence of both symptoms. Based on these foundations, in the performance of ML, the AUC of the classification performance using the LDA classifier model can be as high as 0.74 by 3-fold cross-validation results. The AUC of the results using PDA can be as high as 0.83, both of which have reached the standard of practical application.
Based on the measurement and analysis of changes in arterial pulse wave transmission conditions by observing changes in vascular properties due to sarcopenia, the following conclusions were drawn:
1. Patients with sarcopenia symptoms (PSarco) and general healthy people (Robust) have significant differences in many frequency-domain harmonics.
2. Compared with Robust, subjects with both metabolic syndrome and sarcopenia symptoms (PSarco_PreMetS) had the most significant differences.
3. Using ML analysis, after 3-fold cross-validation results, the classification performance AUC can be as high as 0.74 when using LDA analysis. Compared with PDA, the operation is more careful and comprehensive. As the number of samples continues to increase, it is expected to have better performance. Suitable for cloud computing.
4. Using PDA analysis, the classification performance AUC can be as high as 0.83, and the calculation is simple and fast, which is suitable for device-side calculation.
5. Current measurements and analyses (including ML analysis and PDA analysis) to detect changes in arterial pulse wave transmission conditions caused by sarcopenia help to provide a non-invasive and easy-to-use method to identify possible sarcopenia (PSarco), in the future, it is expected to combine the cloud architecture to provide users with more comprehensive services.

[1] Cruz-Jentoft AJ, Baeyens JP, Bauer JM, et al: Sarcopenia: European consensus on definition and diagnosis: Report of the European Working Group on Sarcopenia in Older People. Age and ageing 2010; 39: 412-23
[2] Tae Nyun Kim, Kyung Mook Choi. Sarcopenia: Definition, Epidermiology, and Pathophsiology. J Bone Metab 2013; 20:1-10.
[3] Rosenberg IH. Sarcopenia: origins and clinical relevance. J Nutr. 1997 May;127(5 Suppl):990S-991S. doi: 10.1093/jn/127.5.990S. PMID: 9164280.
[4] 《肌少症》如何篩檢 & 診斷肌少症？快來看最新版的歐洲肌少症共識！ - 葉峻榳診所
https://www.yehclinic.com/sarcopenia-revised-european-consensus-on-definition-and-diagnosis/
[5] 長庚醫訊│長庚紀念醫院 (cgmh.org.tw)
https://www.cgmh.org.tw/cgmn/category.asp?id_seq=1903012 - .YgIctnpByUk
[6] Chen LK, Woo J, Assantachai P, Auyeung TW, Chou MY, Iijima K, Jang HC, Kang L, Kim M, Kim S, Kojima T, Kuzuya M, Lee JSW, Lee SY, Lee WJ, Lee Y, Liang CK, Lim JY, Lim WS, Peng LN, Sugimoto K, Tanaka T, Won CW, Yamada M, Zhang T, Akishita M, Arai H. Asian Working Group for Sarcopenia: 2019 Consensus Update on Sarcopenia Diagnosis and Treatment. J Am Med Dir Assoc. 2020 Mar;21(3):300-307.e2. doi: 10.1016/j.jamda.2019.12.012. Epub 2020 Feb 4. PMID: 32033882.Cruz-Jentoft AJ, Bahat G, Bauer J, Boirie Y, Bruyère O, Cederholm T, Cooper C, Landi F, Rolland Y, Sayer AA, Schneider SM, Sieber CC, Topinkova E, Vandewoude M, Visser M, Zamboni M; Writing Group for the European Working Group on Sarcopenia in Older People 2 (EWGSOP2), and the Extended Group for EWGSOP2. Sarcopenia: revised European consensus on definition and diagnosis. Age Ageing. 2019 Jan 1;48(1):16-31. doi: 10.1093/ageing/afy169. Erratum in: Age Ageing. 2019 Jul 1;48(4):601. PMID: 30312372; PMCID: PMC6322506.
[7] Kang YJ, Yoo JI, Ha YC. Sarcopenia feature selection and risk prediction using machine learning: A cross-sectional study. Medicine (Baltimore). 2019 Oct;98(43):e17699.
[8] Lenchik L, Boutin RD. Sarcopenia: Beyond Muscle Atrophy and into the New Frontiers of Opportunistic Imaging, Precision Medicine, and Machine Learning. Semin Musculoskelet Radiol. 2018 Jul;22(3):307-322.
[9] Beaudart, C., Rizzoli, R., Bruyère, O. et al. Sarcopenia: burden and challenges for public health. Arch Public Health 72, 45 (2014). https://doi.org/10.1186/2049-3258-72-45
[10] He N, Zhang Y, Zhang L, Zhang S, Ye H. Relationship Between Sarcopenia and Cardiovascular Diseases in the Elderly: An Overview. Front Cardiovasc Med. 2021;8:743710. Published 2021 Dec 9. doi:10.3389/fcvm.2021.743710
[11] Rubio-Ruiz ME, Guarner-Lans V, Pérez-Torres I, Soto ME. Mechanisms Underlying Metabolic Syndrome-Related Sarcopenia and Possible Therapeutic Measures. Int J Mol Sci. 2019;20(3):647. Published 2019 Feb 2. doi:10.3390/ijms20030647
[12] Jung CH, Cho YY, Choi D, Kim BY, Kim CH, Mok JO. Relationship of Sarcopenia with Microcirculation Measured by Skin Perfusion Pressure in Patients with Type 2 Diabetes.Endocrinol Metab (Seoul). 2020;35(3):578-586. doi:10.3803/EnM.2020.679
[13] Pizzimenti M, Meyer A, Charles AL, Giannini M, Chakfé N, Lejay A, Geny B. Sarcopenia and peripheral arterial disease: a systematic review. J Cachexia Sarcopenia Muscle. 2020 Aug;11(4):866-886. doi: 10.1002/jcsm.12587. Epub 2020 Jul 10. PMID: 32648665; PMCID: PMC7432591.
[14] Kalinkovich A, Livshits G. Sarcopenic obesity or obese sarcopenia: a cross talk between age-associated adipose tissue and skeletal muscle inflammation as a main mechanism of the pathogenesis. Ageing Res Rev. (2017) 35:200–21. 10.1016/j.arr.2016.09.008
[15] Li Z, Tong X, Ma Y, Bao T, Yue J. Relationship Between Low Skeletal Muscle Mass and Arteriosclerosis in Western China: A Cross-Sectional Study. Front Cardiovasc Med. 2021 Oct 20;8:735262. doi: 10.3389/fcvm.2021.735262. PMID: 34746254; PMCID: PMC8563701.
[16] O'Rourke MF, Adji A, Safar ME. Structure and Function of Systemic Arteries: Reflections on the Arterial Pulse. Am J Hypertens. 2018; 31(8): 934-940.
[17] Johnson JE, Shay O, Kim C, Liao C. Wearable Millimeter-Wave Device for Contactless Measurement of Arterial Pulses. IEEE Trans Biomed Circuits Syst. 2019; 13(6): 525-1534.
[18] Husmann M, Jacomella V, Thalhammer C, Amann-Vesti BR. Markers of arterial stiffness in peripheral arterial disease. Vasa. 2015; 44(5): 341-8.
[19] Mackenzie IS, Wilkinson IB, Cockcroft JR. Assessment of arterial stiffness in clinical practice. QJM. 2002; 95(2): 67-74.
[20] Davies JI, Struthers AD. Beyond blood pressure: pulse wave analysis--a better way of assessing cardiovascular risk? Future Cardiol. 2005; 1(1): 69-78.
[21] Spectral analysis and dynamics of heart rate and blood pressure in hypertensive patients Victor-Dan Moga 1, Ioana Cotet 1, Andrei Beceanu 1 *, Mariana Moga 2,Florin Vidu 2, Attila Frigy 3, Rodica Avram 1 Romania Received: March 27, 2017, Accepted: April 28, 201 http://hypertens.org/contents/pdfs/jhr-201706-030203.pdf
[22] Applying Pulse Spectrum Analysis to Facilitate the Diagnosis of Coronary Artery DiseaseYi-Chia Huang,1,2 Yu-Hsin Chang , 3 Shu-Meng Cheng,4 Sunny Jui-Shan Lin , 1,2,5,6Chien-Jung Lin,2 and Yi-Chang Su 1,6,7
[23] Bor-Seng-Shu E, Kita WS, Figueiredo EG, Paiva WS, Fonoff ET, Teixeira MJ, Panerai RB. Cerebral hemodynamics: concepts of clinical importance. Arq Neuropsiquiatr 2012; 70(5): 352-356.
[24] Chang YW, Hsiu H, Yang SH, Fang WH, Tsai HC. Characteristics of beat-to-beat photoplethysmography waveform indexes in subjects with metabolic syndrome. Microvasc Res. 2016; 106: 80-87.
[25] Lin SK, Hsiu H, Chen HS, Yang CJ. Classification of patients with Alzheimer’s disease using the arterial pulse spectrum and a multilayer-perceptron analysis. Sci Rep 2021; 11, Article number: 8882.
[26] Lin FL, Hsiu H, Chiu HS, Chen CT, Hsu CH. Characteristics of pulse-waveform and laser-Doppler indices in frozen-shoulder patients. Biomed Signal Process Control 2020; 56: 101718.
[27] Hsiu H*, Liu JC, Yang CJ, Chen HS, Wu MS, Hao WR, Lee KY, Hu CJ, Wang YH, Fang YA. Discrimination of vascular aging using the arterial pulse spectrum and machine-learning analysis. Microvasc Res. 2021; 139: 104240
[28] 翁婉玲（2021）。穿戴式脈波量測系統結合機器學習於失智症及社區場域心智健康狀態評估。國立臺灣科技大學醫學工程研究所碩士論文，台北市。
[29] 楊長仁 （2020）。穿戴式脈波量測系統結合機器學習於社區民眾血管狀態評估。國立臺灣科技大學醫學工程研究所碩士論文，台北市。
[30] 楊舒涵（2015）。都卜勒微循環血流與動脈血壓波形指標 應用於代謝症候群之研究。國立臺灣科技大學醫學工程研究所碩士論文，台北市。
[31] 胡庭毓（2021）。穿施壓對脈波波形品質影響探討及穿戴式裝置應用於肌少症早期偵測。國立臺灣科技大學醫學工程研究所碩士論文，台北市。