胡媛,张江安,孙莉婷,孙钰桢,张涵茵,毛雨,于建斌
[目的]采用聚类分析探讨手部白癜风的临床分型及其意义。方法招募2019年4月-2022年2月在郑州大学第一附属医院皮肤科就诊的手部白癜风患者，并收集患者的临床资料。将手部白癜风划分为8个区域，采用ImageJ计算各区域白斑面积,对白斑的部位及其面积进行聚类分析，聚类分析采用选代分区的k-means聚类,距离算法采用欧式距离法。结果共招募255例手部白癜风患者，其中男181例(70.98%)女74例(2902%)就诊年龄8~68岁(33.57±12.81)岁.发病年龄6~62岁(28.09±1264)岁1，病程~30年557638)年。聚类分析示手部白癜风可分为三类不同亚型，各型在就诊年龄、病程、累及部位(腹部、腕部和下肢)、全身皮损的体表面积(bodysurface areaBSA)差异均有统计学意义(P<0.05):手指型(166例，65.10%)为最常见类型，就诊时较年轻，病程较短，主要累及手指伯侧等易受摩擦部位,BSA<1%为主;手背型(66例，25.88%)次之，就诊时较年轻，病程较短，以累及手背为主，易合并累及腹部(P=0.008)、腕部(P=0.003)及下肢(P<0.01)BSA>5%多于另外两型(P<0.1);手掌型(23 例.9.02%)以累及手学为主，最为罕见，就诊时较年老(P<0.05)，病程较长(D<0.05).BSA在1%~5%之间多见。结论，基于聚类分析，手部白粉风可分为手指型、手掌型和手背型，三类各具有不同临床特点。
[关键词]白癜风;手部;聚类分析;表型
[中图分类号]R758.41 文献标志码]B [1001-7089(2023)03-0291-08
[DOI] 10.13735/j.cjdv.1001-7089.202208007
Cluster Analysis of Clinical Phenotypes in 255 Patients with Hand Vitiligo
HU YuanZHANG Jiang'anSUN Liting.SUN YuzhenZHANG HanyinMAO YuYU Jianbin (Department of Dermatology, the First Afiliated Hospital of Zhengzhou Unwersity,Zhengzhou,
450052，China)
[Corresponding author] ZHANG JianganE-mail:zja816@163.com
[Abstract] Objective To expore the clinical classification and significance of hand vitiligo by cluster analysis. Methods Patients with hand vitiligo who presented to the Department of Dermatology,the First Affiliated Hospital of Zhengzhou University from April 2019 to February 2022 were recruited For all patients their clinical data were collected. Hand vitiligo was divided into 8 regionsImage J was used to calculate the percentage of vitiligo area in each region, and cluster analysis was performed on region and the percentage. K-means clustering with iterative partitioning was used for cluster analysis, and Euclidean distance method was used for distance algorithm.
Results A total of 255 patients with hand vitiligo were recruited,including 181 males(70.98%) and 74 females(29.02%): the age at diagnosis and onset were 868 years(33.57+ 12.81) years and 6-62 years(28.09 ± 12.64)years,respectively, and the duration of disease was 0-30 vears(5.57+ 6.38)vears.Cluster analysis showed that hand vitiligo could be divided into three different subtypes, and for each subtype,there were statistically significant differences in agedisease duration, affected parts(abdomen,wrist and lower extremities) and bodysurface area(BSA) of skin lesions of the whole body(P<0.05) with finger subtype(166 cases.65.10%) as the most common subtype, with younger age, shorter duration of disease, mainly affecting the extensor part of fingersand mainly BSA<1%.The
dorsal hand subtype(66 cases.25.88%) was the second most common, with younger age, shorter duration of disease, mainly involving dorsal hand,and easily involving abdomen(P=0.008) wrist(P=0.003)and lower limbs(P<0.001)and BSA >5% , which was obviously more than the other two subtypes(P<0.001).Palma subtype( 23 cases. 9.02%) mainly affected the palm, which was the rarest,and generally found in patients with older age(P<0.05),longer duration of disease (P<0.05), and BSA of 1%-5%.ConclusionBased on cluster analysis.hand
vitiligo can be divided into finger subtype, palm subtype and dorsal hand subtype, and each of the three subtypes has different clinical characteristics.
[Key words] Vitiligo:Hand:Clusteranalysis;Phenotype
白癜风是一种由黑素细胞选择性破坏引起的皮肤获得性慢性脱色性疾病.影响着全球0.1%~2%的人口1。本病可累及任何部位的皮肤和/或黏膜,包括头面部、躯干、四肢和手足等。发生部位不同对患者的影响也不同，面部、手足等暴露部位对患者影响更大。Florez-Pollack等研究发现手部白癫风与生活质量的显著负面影响相关，且为最难治疗的部位之一。Bae等采用聚类分析将面部白瘢风分为三类，分别为中央型、全面型和发际型，并认为该分型对于探索病因、预测疗效和预后有重要意义。聚类分析是一种数据挖掘方法，将具有相似属性的对象分组为各自的聚类，目前被用于确定某些疾病的特定表型4，但迄今为止，尚无对手部白瘢风的分型。基于此，笔者采用聚类分析的方法对手部白癜风进行临床分型，并探讨其临床意义。
1 对象与方法
11研究对象招募2019年4月-2022年2月在郑州大学第一附属医院皮肤科就诊的手部白癜风患者,并收集患者的临床资料。
111纳人标准手部受累的白癜风患者,符合白癜风诊疗标准(2021年)[5]。
1.12排除标准节段型白癜风患者;②不同意参加本研究者;③精神疾患不能配合信息采集者。
1.2研究方法
121病史信息及临床表型包括性别、就诊年龄、发病年龄、病程、首发部位、其他累及部位、家族史、同形反应及BSA。
122照片收集所有皮损照片均采用佳能EOSD60相机在相同的拍摄部位 角度模式及背景下拍摄,所有拍摄均由同一位医生操作完成。
123面积计算将手部划分为左手指伸侧左手背侧、左手指屈侧、左手掌侧、右手指伸侧、右手背侧、右手指屈侧和右手掌侧共8个区域.采用软件ImageJ计算白斑占所在区域的百分比。
1.2.4纳人聚类分析的变量 白斑的部位和占所在区域的百分比。
12.5聚类分析方法的选择聚类分析采用迭代分区的k-means 聚类,距离算法采用欧式距离法。结合既往文献报道及临床可行性，将全部患者分为2~4类，对各种分类分别进行临床特征的描述后发现分为3类时各类特点比较突出，可以较好地解释异质性,且在临床上易于对患者进行分类操作。
1.3统计学方法应用软件SPSS26.0版(IBMAmonk.USA)进行统计分析。计量资料以x±s表示,符合正态分布采用t检验进行统计分析，不符合采用非参数检验;计数资料采用卡方检验和Fisher 确切概率法。P<0.05 被认为差异有统计学意义。
2 结果
21 临床资料 共筛选出255例患者其中男181例(70.98%),女74例(29.02%),男: 女为2.44:1，就诊年龄8~68岁(33.57±12.81)岁，发病年龄6~62岁(28.09±12.64)岁1.21~30岁(80例，31.37%)多见，病程0~30年(5.57±6.38)年。发病年龄在男性患者[(28.51±12.72)岁和女性患者[(27.07±12.47)岁之间差异无统计学意义(t=0.829 P=0.408):43 例(16.86%)患者家族史阳性,家族史阳性患者发病年龄(24.64+11.99)岁]<家族史阴性患者[(28.79±12.68)岁](t=1.976.P=0.049)。233例(91.37%)患者累及其他部位,其中腕部(139例.54.51%)最常见，其次为头面部(113例44.31%)(图1);BSA>5%组就诊年冷(39.09+13.90)岁 大于BSA<1%组311612.27)岁和BSA1%~5%组[(33.67+12.25)岁(t，= 3.518,P=0.001;[,= 2.408,P,= 0.017),BSA<1%组和BSA 1%~5%组差异无统计学意义(t=1.451,P=0.148);在病程方面,BSA>5%组[(10.07±7.32)年长于BSA<1%组(3.01±4.13)年和1%~5%组[(6.23±6.65)年(=6.003,P<0.001;=3.004,P_=0.011),BSA 在1%~5%组病程同样长于BSA<1%组(t=4.215，P<0.001)(图2)。
2.2 聚类分析 最佳聚类为三类(图3~4)三类手部白癜风特点大致如下(表1):手指型为最常见类型(166例65.10%)，以累及手指伸侧为主就诊时年龄较年轻，病程较短，家族史及同形反应阳性率低，除手部外首发部位以头面部最为多见，易累及腕部,累及3个部位最多见，BSA<1%为主;手背型以累及手背为主,就诊时年龄较年轻,病程较短，家族史阳性率低，同形反应阳性率高，除手部外首发部位以腹部最为多见,易累及腕部及下肢，累及4个部位最多见，BSA在1%~5%之间多见，且BSA>5%明显多于手指型和手掌型:手掌型最为罕见(23 例，9.02%),以累及手掌为主.就诊时较年老(t=2.737,P=0.007;=2.592.P,=0.012).病程较长(t'= 2.187.p=0.038: t',= 2.892.P，=0.008),家族史及同形反应阳性率高除手部外首发部位以腰部最为多见,易累及腕部和足部。累及2个部位最多见BSA在1%~5%为主(图5~7)。
3 讨论
白癜风是一种因表皮黑素细胞破坏引起的损容性皮肤病，发病率逐年上升向。通过招募手部白瘢风255例,本观察首次采用聚类分析的方法将手部白癜风分为手指型手背型和手堂型。手指型为最常见类型(65.10%)主要发生在易受摩擦的手指伸侧部位Gauthier等提出反复的摩擦刺激可能导致易感者的黑素细胞脱离和经表皮消除。此外,研究发现*,肢端白癜风的皮损中肌腱蛋白C(TenascinC)的表达显著升高，其可抑制黑素细胞与纤维连接蛋白的黏附且能直接影响黑色素的生成,在较大的机械应力下使黑素细胞脱落。也有学者认为,较大的机械应力还可引起肢端部位的E-钙黏素(E-cadherin)表达降低导致细胞间黏附丧失和黑素细胞迁移能力增加，从而导致手部白癜风[9]。
手背型主要发生于手部背侧BSA>5%较手指型和手掌型多见。光暴露可能在其中发挥重要作用。Larsabal等[0研究显示，应用PD-1治疗皮肤恶性肿瘤(主要为黑素瘤)的部分患者在光暴露区域出现白癜风样皮损，由多个斑点、斑片演变为更大的斑片，与此类型分布模式相似。此外，Mosenson等1研究表明,光暴露相关的热休克蛋白(HSP70i)可诱导白癜风小鼠模型的树突状细胞表型,为白癫风小鼠脱色所必需，进一步证实光暴露在白癜风发病机制中的作用。
手掌型为一种独特的亚型，以累及手掌为特征，就诊年龄晚、病程较长，最为罕见(9.02%)此型的发生可能与Dkkl(Dickkopfl)在掌跖皮肤中的高表达有关。还因手掌易接触外源性物质，可导致黑素细胞应激.干扰酪氨酸酶和黑色素的合成[2]。另外,掌跖部位黑素细胞和干细胞库密度低，毛囊分布少甚至缺如.发生色素脱失时，此处与正常肤色差异交小,不易引起患者直视就医，从而使此型就诊率远低于正常且延迟诊断。既往的多项研究表明[13-16]白癜风多见于女性，而在Gandhi等[]和自翠彦等[8]学者的研究中,男性略多于女性。在本研究中,手部白癜风男性远多于女性,男女比例2.44:1.且20岁之前发病患者仅占30.58%，低于既往报道的50%[10]。手部白癜风多见于男性且发病年龄偏晚,可能归因于成年男性较女性从事更多的体力劳动,增加了手部摩擦受损的机会，导致本病的发生。值得注意的是,日常生活中腕部易受縻擦,本研究发现手部白瘢风极易合并累及腕部(139 例。54.51%),但腕部为首发部位极为罕见(2 例，0.78%)，提示腕部白癜风可能存在重为复杂的机制。此外,本研究中43例(16.86%)患者家族史阳性,家族史阳性患者发病年龄[(24.64199)岁早于家族史阴性患者[(28.79+12.68)岁](t=1.976,P=0.049)，与Karelson等研究一致提示遗传易感性在其中发挥重要作用。Surekha等[20]分析3551 例印度白瘾风患者后发现，X连锁基因和常染色休基因可能共同参与本病的发生，在一定程度上可解释本研究手部白瘾风以男性多见。
本研究存在一定的局限性。首先，未对疗效进行统计，从而评估各型间疗效是否具有差异性。其次,临床中发现，极少部分患者几乎累及手部所有部位,应用上述三类分型不能完全解释其临床特点，提示其可能为一独特亚型,需后续进一步的研究。综上所述，笔者利用聚类分析将手部白癌风分为三个亚型:手指型白癜风为最常见的类型;手背型为一独特亚型,除腕部外,易累及腹部和下肢，BSA>5%较另外两型多见:手掌型为另一种独特的亚型,最为罕见，就诊年龄大，病程长.临床中易被忽视。三类不同的临床表现提示背后存在不同的发病机制,本研究的分类将有助于探索病因、预测疗效和预后。
[参考文献]
[1] Bergqvist CEzedine K.Vitiligo:A review[J].Derma-tology， 2020, 236(6): 571 -592.
[2] Florez-Pollack S,Jia G,Zapata L, et al. Association ofquality of life and location of lesions in patients with vitiligo[J].JAMA Dermatol，2017153(3):341-342.
[3] Bae JM,Jung YSJung HM,et al.Classification of facialvitiligo:A cluster analysis of 473patients[J].Pigment Cell Melanoma Res2018,31(5):585-591.
[4]Jung S,Moon J.Hwang E.Cluster-hased analysis of infec-tious disease occurrences using tensor decomposition: Acase study of South Koma[J].Int J Environ Res Public Health，2020，17(13):4872.
[5]许爱娥白癜风诊疗共识(2021版)[J].中华皮肤科杂志,2021,54(2):105-109.
[6] Fzzedine KEleftheriadou VWhitton Met al.Vitiligo[J].Lancet，2015,386(9988):74-84.
[7] Gauthier Y,Cao-AndreMLepreuxSet al.Melanocyte detachment after skin friction in non lesional skin of patients with generalized vitiligo[J].Br J Dermatol,2003,148(1):95-101.
[8] Esmat SM,Hadidi HHE,Hegazy RA,et al.Increased tenascin C and DKK1 in vitiligo:possible role of fibroblasts in acral and non-aeral diseasej1. Arch Dermatol Res.2018 310(5):425-430.
[9]Levy C,Khaled M. Ecad vitiligone[J].Pigment Cell Mel-anoma Res，201528(4):376-377.
[10] Larsahal M，Marti A,Jacquemin C，et al.Vitiligo-like lesions occurring in patients receiving anti-programmed cell death-1 therapies are clinically and biologically distinct from vitiligo[J].J Am Acad Dermatol201776(5)63-870.
[11] Mosenson JA,Zloza A,Nieland JD,et al.Mutant HSP70 reverses autoimmune depigmentation in vitiligo[J].Sci Transl Med，20135(174):174ra28.
[12] Rodrigues M,Ezzedine K,Hamzavi Iet al.Vitiligo Working Group. New discoveries in the pathogenesis and classification of vitiligo[J].J Am Acad Dermatol,2017，77(1):1-13.
[13] Kussainova AKassymL，Akhmetova A,et al.Vitiligo and anxiety:A systematie review and meta-analysis[J].PLoS One, 202015(11): e0241445.
[14] Canu D,Shourick J,Andreu N，et al.Demographic and clinical characteristics of patients with both psoriasis and vitiligo in a cohort ofvitiligopatients: a cross-sectional study[J].J Eur Acad Dermatol Venereol，2021，35(10):e676-e679.
[15]Zhang L,Chen S,Kang Y,et al.Association of clinical markers with disease progression in patients with vitiligo from China[J].JAMA Dermatol2020156(3):288-295.
[16] Hadi A,Wang JF,UppalPet al.Comorbid diseases of vitil igo: A 10-year cross-sectional retrospective study of an urhan US population[J].J Am Acad Dematol202082 (3):628-633.
[17] Gandhi KEzzedine KAnastassopoulos KP,et al.Pre-valence of vitiligo among adults in the United States[J].JAMA Dermatol，2022158(1):43-50.
[18]闫翠彦，张峻岭。索丹凤，白癜风患者 406 例临床调查分析[J]，中国皮肤性病学杂志，2010，24(2);123-125.
[19] Karelson M,Silm H,Salum Tet al.Diflerences between familial and sporadic cases ofvitiligo[J].J Eur Acad Dennatol Venereol2012 26(7):915-918.
[20] Surekha TIshaq M,atha KP,et al.Do clinical vari-ants of vitiligo involve X-linked gene(s)too?[J]. JMed Sci2008.8:728-733.
[9]Mevorach D,Mascarenhas JO,Gershov Det al.Comple-ment-dependent clearance of apoptotic cells by human mac-rophages[J].JExp Med1998188(12):2313-2320
[10] Fraser DA,Laust AKNelson ELet al.Ciq differentially modulates phagoeytosis and cytokine responses during ingestion of apoptotic cells by human monocytes,macro-phages, and dendritic cells[J]. J Immunol,2009,183(10):6175-6185.
[11] Martin M,Blom AM.Complement in removal of the dead.balancing inflammation[J]. Immunol Rev，2016，274(1):218-232.
[12] Sereflican B,Bugdayci G.Components of the altemative complement pathway in patients with psoriasis[J].Acta Denmatovenerol Alp PannonicaAdriat,2017,26(2):37-40.
[13] Ricklin D,Reis ES,Mastellos DC，et al.Complement component C3-The "swiss army knife" of innate immuni- ty and hast defense[J].Immunol Rev,2016,274(1):33-58.
[14] Purwar RWittmann M,Zwirner J,et al.Induction of C3 and CC12 by C3a in keratinocytes: a novel autocrine amplification loop of inflammatory skin reactions[J].J Immunol 2006.177(7):4444-4450.
[15] Purwar R,Baumer WNiebuhr Met al.A protective role of complement component 3 in T cell-mediated skin inflammation[J]. Exp Dermatol，201120(9):709 -714.
[16] O'Brien KB.Momison TEDundore DYet al.A potec-tive role for complement C3 protein during pandemic 2009 HIN1 and H5NI infuenza A virus infectionj1.PLoS One,2011,6(3):e17377.
[17]Giacomassi C,Buang NLing set al.Complement C3 exacerhates imiquimod-induced skin inflammation and psoriasiform dermatitis[J].JInvest Dermatol,2017137(3):760-763.
