Characteristics of mucocutaneous vascular malformations drawn from a decade of a multidisciplinary committee experience

Vascular malformations (VM) are congenital, benign, and relatively frequent lesions. Scant data have been published about the epidemiology, clinical presentation, and treatment of VM from a dermatologist's perspective. The substantial differences between subtypes, broad range of specialists consulted and confusing nomenclature used over previous years may hamper a correct diagnosis. The main objective of this study is to describe VM epidemiology. As a secondary endpoint we evaluate clinical characteristics, clinical‐radiological correlation and treatment approaches. We carried out an observational, descriptive, retrospective study. Cases presented to the multidisciplinary committee of our hospital from 2009 to 2019 were retrieved. Electronic medical records, monthly committee reports and the iconographic archive were reviewed and statistically analyzed. Overall, venous malformations (VeM) are the most frequent VM, followed by capillary malformations (CM), arterioVeM and lymphatic malformations (LM). Considering only patients under 16, CMs are the most frequent ones. Capillary and LMs are larger than venous or arteriovenous. While CMs are usually asymptomatic, symptomatic cases are threefold more frequent in the other subtypes. Decisions on active or conservative management depend on VM size but not location or patient age. CMs are mainly treated with laser therapy; venous with sclerotherapy or surgery; arteriovenous with surgery and lymphatic with surgery or sirolimus. Dermatologists play an important role in VM diagnosis and management. Our 10‐year multidisciplinary experience should contribute to the literature and represent a practical resource for clinicians and researchers.


| INTRODUCTION
Vascular anomalies are divided into vascular tumors and vascular malformations (VM). VM are benign lesions derived from a vasculogenesis failure, without endothelial proliferation, unlike vascular tumors.
Although congenital, they may not become evident until adulthood.
VM comprise a wide range of entities, and current terminological confusion is partly attributable to previous classifications. The International Society for Vascular Anomalies (ISSVA) was founded in 1992, and from this year onwards classification of vascular anomalies has been reviewed every 2 years, with its latest update in May 2018 1 and a recently proposed new clinical classification for capillary malformations (CM). 2 Here we report the epidemiological, clinical, radiological and therapeutic characteristics of more than 200 VM evaluated in our hospital over 10 years, exploring differences between the main subtypes to help advance knowledge in the field.

| VM subtypes
VM are due to mosaic mutations, most of them involving the Ras-MAPkinase or PI3K pathways. VM are divided into simple (CM, venous malformations (VeM), lymphatic malformations (LM), and arterioVeMs (AVM)) and combined if two or more type of vessels are involved.

| Capillary malformations
These malformations have seven major patterns which differ in clinical presentation, prognosis and possible associated findings. 2 Nevus simplex or flammeus usually presents as pale pink to bright red macules located in middle face or head. 2 Port-wine stains present as pink, red or purple stains varying in size; however, the most common presentation is large, unilateral well demarcated patches with segmental distribution.
Geographic CM are typically blue to purple, well demarcated stains, often associated with veno-LMs and limb overgrowth, such as in Klippel-Trenaunay syndrome, CLOVES, Proteus or CLAPO syndromes.
Small multiple CM are usually associated with CM-arterioVeM syndrome. 3 Reticulated CM are characterized by reticulated, poorly defined, pale pink to light red vascular stains. These lesions can be isolated or associated with overgrowth such as in diffuse CM with overgrowth syndrome or in Macrocephaly-CM syndrome. Cutis marmorata telangiectatica congenita is characterized by a congenital stain with blue to purple tone and a livedoid pattern with frequent focal atrophy or ulceration. Telangiectasia is a very heterogeneous group characterized by dilated capillary vessels. Although, according to ISSVA, telangiectasias are classified as CMs, most of them (such as in Rendu Osler Weber syndrome) present as high flow lesions. 4

| Venous malformations
VeM are rare congenital low-flow VMs with a prevalence of 1%. 5 Clinically they present as soft nodules or masses with a blue to purple color and increase in size with the Valsalva maneuver. Pain (secondary to thrombosis or nervous compression), functional limitations and bleeding are common. There are several different subtypes.

| ArterioVeMs
AVM are fast-flow VM caused by an abnormal connection between artery and vein, which represent between 10% and 15% of all VM. 6 Typically, they present as reddish macules or patches similar to CM but with fast-flow Doppler vessels (stage I Schobinger's clinical classification 6 ). As they evolve, they present with thrill, ulceration, bleeding, pain or necrosis or even heart failure. The reason why some AVM stay stable while others quickly progress is not clear, although genetics could play an important role.

| METHODS
We carried out an observational, descriptive, retrospective study. We collected VM patient cases discussed between All data were statistically analyzed, using the nonparametrical Pearson's chi-squared (Chi2) test in the case of two categorical variables and performing Bonferroni correction for Post hoc analysis for more than two categories in any variable. To determine whether data distribution among different categories (such as sex, size, or tone) was balanced, we performed Pearson's Chi2 test for a single variable. To compare a continuous variable with nonnormal distribution or an ordinal variable we used different tests depending on the second variable: for categorical variables with two possible categories, we used Mann Whitney U test, while the Kruskall-Wallis test was used for categorical variables with more than two categories.
To compare two continuous variables with nonnormal distribution we used Spearman's rank correlation coefficient (ρ), and to evaluate the concordance level between ultrasound and MRI results we used Cohen's kappa coefficient (κ).

| Subtypes of VM
In total, 202 VM cases were included in our study, of which VeM were the most frequent (38%), followed by CM (32%), AVM (23%) and LM (7%). We also added five patients with malformations of major named vessels. Combined malformations were classified according to their main vascular component.

| Sex distribution
Overall, we found a nonstatistically significant difference between VM subtype and patient sex (p = 0.605), although in the fast-flow VM subtype females outnumbered males twofold (p = 0.011) ( Table 1).
T A B L E 1 Sex, size, and location according to VM subtypes VM subtype

| Location
Anatomical distribution by subtype is represented in Figure 2. Statistically significant differences were found between head-neck and trunk location and VM subtype (p = 0.035 and p < 0.0001 respectively) ( Table 1). For VM located in head and neck, the most frequent subtype was VeM, whereas in trunk the most common were CM and LM.

| Color
The color distribution for each subtype is represented in Figure 3. We found differences between VM subtypes in terms of color (p < 0.001).
F I G U R E

| Signs and symptoms
There are differences between clinical manifestations and subtypes of VM (p < 0.0001), as CM were tenfold less likely to be symptomatic than the other subtypes.
• In VeM, symptomatic cases represented 81% of total. Pain (43%), bleeding (12%) and frequent changes in size or volume (47%) were the most related symptoms. Coagulation disorders (D-dimer >300 ng/ml) were present in 8% of VeM, with ultrasoundconfirmed deep vein thrombosis in 50% of those cases. All cases with coagulation disorders but no thrombosis were extensive VeM affecting joint mobility.

| Treatment
Overall, 88 patients were managed conservatively, six with mTOR inhibitors, 19 with sclerotherapy, nine with embolization, 42 with surgery and 49 with laser therapy. Some patients received more than one treatment.
We found no relationship between decision-making on active treatment and VM subtype (p = 0.491), patient age (p = 0.055) or location of VM (p > 0.05 for all categories). However, treatment decision-making was influenced by VM size, as those <5 cm2 were more frequently managed conservatively (p = 0.024).
The management of each subtype is detailed as follows: • CM: 64% of CM were treated with laser therapy (93% PDL and 7% PDL-NdYAG). The mean number of sessions was 6.57 for treatments without sedation and 4.12 for treatments under sedation.
The number of laser therapy sessions was higher for CM ≥5cm 2 (p = 0.015). However, we found a nonstatistically significant difference between number of laser therapy sessions and age of treatment onset (p = 0.024, p = 0.882) or location of CM (p > 0.05 for all categories). We found a trend towards a greater number of sessions in CM located in the lower limbs (p = 0.099).
• LM: 36% of cases underwent surgery and 29% received mTOR inhibitors. All LM treated with oral Sirolimus were large, complicated malformations leading to functional limitation or with laryngopharyngeal involvement.

| Multivariate analysis
Multivariate analysis revealed clinical symptoms, trunk location and age of onset as the best predictors of VM type, dwarfing all other factors studied. With these three categories correct diagnosis of CM and VeM reaches over 85%, of LM around 65% and 53% for AVM.

| DISCUSSION
The distribution of VM subtypes in our study is similar to those Small CM located in nonvisible areas rarely prompt pediatrician referral to tertiary hospitals. However, in our study, AVM were more frequent than in other studies. 6,7 This could be due to several factors: we included all stages (about 70% were stage I-II) and all AVM were studied with ultrasound. As previously mentioned, correlation between different imaging exams was much weaker for AVM than for the other subtypes, which could lead to overdiagnosis.
Regarding location, over 50% of CM, VeM and AVM are located in the head and neck. Compared with previous reports (40%) our study found an even higher proportion of VeM located in head and neck (67.6%), frequently involving the oral mucosa. 8-10 LM can be located anywhere in the body but tend to affect areas with most lymphatic drainage, such as the neck or axilla 9 (Figure 1).  23 and has already been reported in BRB 24 and verrucous VeM. 25 We treated two VeM with oral sirolimus: a combined capillary-VeM and an extensive symptomatic common VeM.
Although surgery with preoperative embolization is considered the standard treatment for AVM, it is a complex procedure requiring a trained team. Oral MEK inhibitors (EudraCT 2019-003573-26) and thalidomide 26 have been proposed to treat inoperable AVM.
Limitations of our study include that the population may not be representative as we are a tertiary hospital where patients are referred for specialist opinion. As it is a retrospective study some data may not have been available; however, detailed medical records of all patients discussed in the committee were at our disposal.

| CONCLUSION
VM is a frequent reason for consultation in dermatology; however, diagnosis and treatment require multidisciplinary management. Our report on the epidemiological, clinical, radiological and therapeutic characteristics of more than 200 VM evaluated in our hospital should contribute to the literature and provide practical data for clinicians and researchers. Genetics will probably play a major role in the future classification, diagnosis and targeted treatment of VM.

AUTHORS CONTRIBUTION
All authors are involved in the multidisciplinary committee according to their medical specialty. Dr Martín Hernández suggested and conceived the study. Dr Martín Hernández, Dr Puche-Torres and Dr Sanchis directed the project. Dr Estébanez carried out the study and wrote the article. All authors discussed the results and contributed to the final version.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.