Synovial haemangioma (SH) is a rare benign tumor accounting for 0.07% of all soft tissue tumors and 0.78% of all excised haemangiomas (1). While this type of lesion is usually monoarticular (2), more than one mass can occur in different locations within the joint (3). SH has been reported in many joints but is most common in the knee (4).

Knee synovial haemangioma (KSH) can present as diffuse or localized intra-articular, juxta-articular or intermediate (intra- and juxta-articular) synovial tissue masses (2, 5–7). Extra-synovial and intra-synovial locations have been described for joint tumors (8), with intra-synovial locations being more likely to affect cartilage (8). It can be a part of an angiodysplastic lesion of the extremities (Klippel-Trénaunay syndrome) with diffuse (9, 10) or localized (11) involvement of the joint.

Most are located in the anterior compartment of the knee (7, 12). Only eight localized cases have been reported in the posterior compartment [one intermediate (5), three juxta-articular (1, 12) and four intra-articular (13–17)], one of which was reported twice (14, 15). Most reports are single case studies (1–3, 11, 13–50), while case series are less common (4–10, 12, 51–57).

The clinical presentation is non-specific, with symptom onset usually occurring in childhood and young adulthood, although it has been described at later ages, probably due to delayed diagnosis. Various imaging modalities have been used, but all are non-specific, and there is no clear diagnostic algorithm (2, 4, 9, 13, 18, 25, 28–30, 32, 33, 38, 45, 48, 50, 55, 57, 58).

Once a benign synovial lesion is suspected, it is unsettled if a biopsy is needed or if excisional biopsy and regional synovectomy with histological examination can confirm the diagnosis and simultaneously accomplish definitive treatment. Whether open (2–5, 8–10, 16, 21, 37, 38, 57) surgery is preferable to arthroscopic (6, 38, 39) surgery remains controversial, especially in localized intra-articular lesions. There is no clear information on the recurrence rate after surgical treatment.

Further investigation is warranted, particularly because the existing information is scattered and based mostly on isolated clinical cases. An understanding of the differences in patient characteristics and clinical presentation, along with knowledge of the real value of imaging examinations and follow-up results after arthroscopic or open surgery, is important for making informed decisions on the need for invasive imaging examinations and on the differential effectiveness of open vs. arthroscopic treatment in light of the evidence.

The aim of the study was to describe four patients with intra-articular KSH, two with the previously undescribed phenomenon of patellar bone erosion and one with an unusual tumor location in the posterior compartment, and to perform a systematic review of PubMed publications on KSH from 2000 to 2020 to describe the characteristics of the patients (age, gender, knee side, previous trauma), the patterns of tumor location, the non-specific clinical presentation, the usefulness of imaging examinations, the pros and cons of arthroscopic vs. open resection, and the follow-up time and recurrence after treatment.

Results tabulated by age, gender, affected knee side, tumor location, history of previous trauma, clinical symptoms and signs, presence of cutaneous haemangioma, time in diagnosis delay, imaging test used, bone involvement, type of treatment, histology, follow-up time and recurrence are shown in Table 1 for the clinical cases and Table 2 for the literature review.

Synovial haemangioma have been described in various joints, but the knee is the most common (4). Up to 2015, one paper reported more than 275 cases of KSH (1). Since then, another 24 cases have been reported (PubMed only), with the present four cases bringing the total to 303. There are a number of small case series (4–10, 12, 51–57, 61), the largest describing 14 cases (8). Most reports describe single clinical cases (1–3, 11, 13–50), which highlights the rarity of this type of tumor.

This is the first study that performs a detailed revision of a large number of KSH (96 cases) and describes its more frequent location and clinical presentation forms. The major finding of this study is that plain radiographs and MRI image studies appear to be enough for the differential diagnosis, that will be confirmed after histology exam. Being a tumor mass of benign characteristics on image studies rarely a biopsy is indicated previous to treatment, an excisional biopsy and regional synovectomy can be performed for diagnosis and treatment purposes before diagnostic confirmation. Although open surgery has been the treatment of choice, nowadays arthroscopy seems to offer advantages in the treatment of intraarticular KSH without bone, tendon or muscle involvement. Recurrence after surgery is low (3.1%) but persistence of symptoms is more frequent (11.4%), especially in diffuse forms.

In the last 20 years, only nine cases (9.4%), including ours, had localized lesions in the posterior compartment of the knee. One was an intermediate tumor (5) that extended into the quadriceps and gastrocnemius. Three were juxta-articular, one with an intramuscular component in the popliteal fossa (1) and two behind the posterior cruciate ligament (PCL) (12). The other five were intra-articular, of which three previously reported cases plus our case 4 were behind the PCL (13–16) and one was just cranial to the PCL (17).

Our other three cases were local sessile intra-articular SH located near the superior pole of the patella, by far the most frequent location in the anterior compartment of the knee (5–8, 19, 21, 24, 27–29, 38, 40, 42, 44, 50, 53–57), followed by the infrapatellar fat pad (11, 18, 23, 30, 45, 59), the posterior compartment, the medial compartment (2–4, 25, 35, 39, 42, 54, 61) and the intercondylar notch (22, 43, 46, 47, 54). The three reported patients with KSH near the superior pole of the patella had the rare phenomenon of bone erosion, which was localized in the patella in two cases, a previously undescribed location for KSH bone erosion, and in the supratrochlear cortical bone in one case. It is most frequent in tumors located in the medial compartment (2, 3, 6, 39) or diffusely distributed (31, 34).

KSH usually occurs in children and teenagers (4, 8, 12), although the average age is 21.7 ± 14.7 years. For that reason, the diagnosis is not typically suspected in middle-aged or older adults, as occurred in our patients. The average delay in diagnosis was 6.8 ± 8.6 years, larger than the average noted in one report (6), although two reported patients had a delay in diagnosis of more than 40 years (3, 41).

All four of our patients were male, but we found no differences in sex distribution, and the exclusively male composition of our sample is probably merely a matter of sample size. The right knee is more frequently affected than the left knee. Previous trauma appears to be rare and is often not reported. Symptoms and signs associated with SH of the knee are non-specific. Low-intensity and long-duration pain is the most frequent symptom (4), but it rarely limits activities of daily living except in phases of exacerbation (5, 12, 59). It is usually aggravated with activity and relieved by rest (2, 3, 5, 10, 42, 49). Pain is typically in the anterior aspect of the knee but can be posterior (14) or diffuse. Swelling, tenderness, effusion (5, 9–11, 17, 21, 23, 26, 30, 33, 34, 44, 49, 50, 55) and haemarthrosis (5, 9, 11, 12, 18, 22, 30, 31, 37, 39, 41, 46, 50, 54, 55, 57, 59) are frequent and can be recurrent. Occasionally, a palpable soft tissue mass (3, 4, 10, 44, 46, 55, 61) can be noted. Anterior masses are more evident in complete flexion of the knee (3, 21, 44) and can be compressible, refilling when pressure is released (21). Limitation of motion (3, 5, 8–10, 12, 17, 25, 28, 33, 37, 38, 46, 61) and mild muscular atrophy (5, 11, 33, 34, 36, 49, 55) can be present. Joint locking is infrequent but has been described when synovial haemangioma is located in the intercondylar notch (22, 47) or in the suprapatellar pouch (6). Although up to 40% of patients with synovial haemangioma may have cutaneous haemangioma (4), we appreciate that it is less frequent, probably because patients are often not carefully examined for such haemangiomas, or, if so, they are not reported. Cutaneous haemangioma has been described in patients with diffuse extra-articular KSH (6, 29, 31, 49) or with systemic angiodysplasia (Klippel-Trénaunay) (22, 23).

No clinical test or imaging procedures provide a definitive diagnosis, nor is there a clear algorithm for diagnosis. Plain radiographs are typically normal (4, 8, 11, 13, 16, 17, 21, 27, 28, 35, 38, 39, 41, 44, 47, 49, 50, 53, 55– 57), and only a few cases show non-specific abnormalities (2, 4–10, 33, 36, 45, 53) such as periosteal reaction, osteolysis, osteopenia, intra-articular phleboliths, trochlear dysplasia or osteoarthritis. Even if there is a small region of bone erosion, as in our cases, plain radiographs can appear normal. Formerly, it was the first imaging test ordered because it is fast and inexpensive and permits a general evaluation of the joint, although it is not useful in the diagnosis of KSH.

MRI appears to be the examination of choice (1, 2, 5–11, 16–18, 21–24, 26, 27, 33, 35–40, 42, 43, 46, 48, 50, 53–57). It is non-invasive and non-irradiating and usually allows three-dimensional localization, although occasionally, as in our series, a tumor mass may not be identified (3, 47). However, the features of low or intermediate intensity compared to muscles on T1-weighted sequences and high intensity on T2-weighted or fat-suppressed sequences are not specific to haemangioma (28, 58). Fatty fibrous septa can be seen within the lesion in some cases (57). Large venous haemangiomata can cause high signal intensity on T1 due to slow blood flow (58). Two reports suggested that only 22% of the cases had a correct suspected diagnosis before surgery (2, 10). MRI is useful for demonstrating extension into the periarticular tissues (7), for detecting bone erosion in our cases, for surgical planning, and for following recurrence. Intravenous administration of gadolinium for MRGd produces inhomogeneous enhancement of the synovial tumor image (9, 14, 27, 34, 38, 46, 59) and permits differentiation between cystic synovial hyperplasia (52) and liquid or intra-articular haemangioma (2, 7, 38, 57). MRA and MDCTA have replaced arteriography, and while both are invasive and non-specific methods (20, 50), they may be justified for planning surgery in diffuse or extended cases (3, 32, 46) and in cases where complete excision is not possible and where embolization might be considered (7, 57).

Ultrasound may (19, 23, 24, 46) or may not (8) help distinguish solid vs. liquid masses or those with heterogeneous content. Echo Doppler may suggest a vascular lesion (23, 38, 46, 50, 60), but not always (28). CT is useful in determining the presence and extent of bone involvement (2, 7, 33, 55). Tc99 scintigraphy may show increased uptake in cases where bone is affected (2, 45), RBCSc reveals increased blood pool activity (30), and SPECT/CT shows a radioactive mass (30); however, the information provided is non-specific, so a scan would be justified only in a patient where other imaging examinations are normal.

Diagnostic arthroscopy may be useful in localizing a synovial tumor and ruling out unrelated intra-articular lesions (6, 33, 45, 47). Arthroscopy allows lesion localization; biopsies for frozen section (16, 55), permanent histology or microbiological study; and, in most patients, definitive treatment. We indicate it in the event of persistent knee pain and normal imaging examinations. However, as in case 2, it is not always possible to visualize the tumor (17, 20, 28, 31, 57, 59, 60) or show the complete extent of the lesion (8), either because it is not localized (59) or misdiagnosed (28) or because the lesion is subjacent to the synovial lining and cannot be seen (20).

Histological examination is essential to establish a definitive diagnosis, although it was not reported in 8.3% of reviewed cases. Haemangiomata are tumors derived from endothelial cells, and four types have been described (51). Although three of our cases were of the arteriovenous type, we showed that the most common was the cavernous type, followed by the venous vascular malformation, capillary, arteriovenous and cavernous-capillary types. In 19.8% of the cases, the histological type was not described. It is considered a vascular venous malformation more than a tumor, often associated with a lymphatic hamartomatous component (8).

We advise treating these lesions because they can cause joint destruction either by direct invasion of the surrounding tissues or by repeated haemarthrosis (5, 10, 31). Embolization may be indicated in cases of diffuse lesions, when complete surgical resection is difficult or impossible, or when the tumor mass is nourished by a large blood vessel (7, 36). Non-surgical treatment and observation have been recommended by some authors in diffuse cases (6, 32, 49).

As KSH appears as a synovial tumor mass with benign characteristics in imaging studies, an excisional biopsy and regional synovectomy can be performed until free resection margins are obtained for diagnosis and treatment purposes before diagnostic confirmation. Biopsy prior to treatment is rarely reported (9.4%) but should be indicated if there is doubt regarding the benignity of the tumor based on imaging examinations. There is controversy as to whether open or arthroscopic surgery is better. Most patients have been treated by open surgery because it has been argued that it allows a better view of the full extent of the lesion, mainly when it is located within the Hoffa fat pad, and therefore leads to a lower recurrence rate (4, 8–10, 18, 20, 36, 50, 55). However, recurrence can occur despite complete resection by open surgery (4, 10). Our study found a 3.1% reported recurrence rate with a 11.4% persistence of symptoms rate. We found no large studies reporting recurrence after open vs. arthroscopic resections of localized or diffuse SH, and the reported follow-up times were short or not reported, but PVNS recurrence rates were reportedly comparable between open and arthroscopic surgery (62). We believe open surgery is warranted with juxta-articular or intermediate tumors (50), that is when the tumor extents to the tendons, muscle or bone.

Some authors recommend arthroscopic resection for localized intra-articular KSH (6, 14, 15, 20–22, 33, 36, 38–40, 42, 43, 47, 56, 57, 59, 60). Hoffa fat pad SH can be arthroscopically resected using suprapatellar accessory portals (11, 20, 59). In the past, posterior knee compartment SH was typically treated by open surgery (1, 5, 12, 16, 17) or non-surgically due to predicted difficulty in surgical resection (13); only one case, reported twice, was treated by arthroscopic surgery (14, 15). We performed arthroscopic resection of a posterior compartment SH with resection instruments in the posteromedial and arthroscope in the anterolateral or anteromedial portal (15, 63). This technique avoids possible complications related to an accessory posterolateral portal but has a limited field of view (64). An arthroscopic trans-septal technique using posteromedial and lateral portals may be useful when more working room is needed in the posterior compartment (14, 15, 64). Arthroscopic surgery is advantageous in that it is a minimally invasive technique allowing good assessment of all the joint compartments; it can be used to rule out and treat associated lesions or tumors in the joint; it promotes rapid recovery and superior cosmesis; and it reduces the incidence of postoperative stiffness, infection, thromboembolic events and joint pain (15, 63). An open posterior approach requires a large incision, preventing examination of other parts of the joint, and is associated with relatively poor cosmesis and slow recovery. With the haemostasis achieved by the use of electrocoagulation during arthroscopic tumor resection and synovectomy, open surgery has no advantage in bleeding control (60).

We note several limitations of our study. First, some clinical data were missing in our retrospective clinical review, although key data were available in the records. Second, the literature search was performed only in PubMed and was limited to English, French or Spanish; thus, there is likely additional literature on the subject. However, given the long period over which the papers were published and the large number of cases identified, the included studies are likely representative of the literature. We are unaware of any other such reviews at this level of detail.

It is clinically relevant to know the frequent and unusual locations of KSH, its multiple clinical presentations, the real value of image examinations and the pros and cons of arthroscopic vs. open surgery to be able to make informed decisions on the need for invasive image examinations, previous biopsy or arthroscopic vs. open surgery in light of the evidence.

In conclusion, KSH is a rare tumor. The superior pole of the patella is the most frequent location and only five intra-articular reported cases (including ours) in the posterior knee compartment. SH should be considered in the differential diagnosis of long-term knee pain in adults, particularly if there is no history of previous trauma. Plain radiography is usually normal. MRI is the imaging test of choice to localize a synovial mass, but the images are diagnostically inconclusive. MRA may be useful in diffuse or extended cases. The rare phenomenon of bone erosion, first time described in the patella, have to be bear in mind. Histological examination is necessary for definitive diagnosis. Biopsy prior to treatment should only be indicated if there is doubt regarding the benignity of the tumor. We believe arthroscopic resection should be the surgical technique of choice for treating localized intra-articular lesions or those in difficult areas such as the posterior compartment. We recommend open surgery for diffuse tumors and those that are localized in juxta-articular and intermediate regions with tendon, muscle or bone extension. Complete resection of the synovial lesion and the surrounding affected tissues is important to relieve symptoms and prevent recurrence.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

JH-H participated in the design of the study, participated in revision of clinical cases and search and revision of bibliography, and interpretation of the data and in drafting the manuscript. JM-B participated in revision of clinical cases and revision of bibliography. EG-O has participated in search and revision of bibliography. FC-S has been involved in search and revision of bibliography. SL-M has participated in search and revision of bibliography, interpretation of data, and drafting the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.