Two researchers independently searched the PubMed, Embase, WanFang, and CNKI databases from the inception to May 2021 to find published articles related to LCA. The search terms used included littoral cell angioma or littoral cell tumor or littoral cell without restrictions. Furthermore, the reference lists of the retrieved studies and recent reviews were reviewed to identify additional potentially relevant articles. This study was approved by the Institutional Review Board of the First Affiliated Hospital of Zhengzhou University (2017-xy-002). The typical imaging and pathological pictures presented in the figures were obtained from our center.

Two investigators independently evaluated all records by screening the title and abstract for potentially eligible studies, and differences in opinion were resolved by consensus with a third reviewer. Studies (i) in which a postoperative pathology confirmed the diagnosis of LCA; (ii) that reported the epidemiology, natural history, pathogenesis, clinical manifestations, treatment and prognosis of LCA; (iii) that were published in English or Chinese before May 2021; were included in the analysis. However, reviews and LCA patients repeatedly published by the same unit were excluded. The clinical information on the epidemiology, pathogenesis, clinical manifestation, treatment, and prognosis of LCA patients were extracted and analyzed.

The graph was drawn using Graph Pad statistics software (Graph Pad Software Inc, USA). The statistical indexes were expressed as mean and standard deviation with enumeration data, which were consistent with the normal distribution. Otherwise, it was expressed as median and interquartile range (IQR).

A total of 189 studies comprising 435 patients met the inclusion criteria ( Figure 1 , Table 1 , and Table S1 ), namely, 171 foreigners and 264 Chinese individuals. The statistical analysis showed a male-to-female ratio of 0.90 in 405 patients with LCA, which indicated the absence of gender predilection. The average age of 267 patients was 48.4 ± 16 years, with the oldest and youngest being 86 years old and 26 days old, respectively (11, 19). As shown in Figure 2 , LCA was more prevalent in the 40–60-year age group; however, it was detected in 15 minors (1, 19–30).

A total of 60 various malignant tumors co-occurred with LCA (1, 11, 12, 31–65), as shown in Table 1 and Figure 2 , with a comorbidity rate of 13.8%. A total of 59 patients had comorbid malignancies, which occurred before the formation of LCA. Only one patient died of gastric cancer 2 years after undergoing splenectomy (53). Considering the typical survival time of gastric cancer, we speculated that this LCA patient was very likely to have developed gastric cancer at or before the time of splenectomy. Moreover, 53 immune disorders occurred along with LCA (1, 9, 11, 24, 27, 34, 35, 39, 44, 48, 50, 52, 66–88), with a comorbidity rate of 12.2%. It is worth noting that 9 patients had both malignant tumors and immune dysfunctions prior to the development of LCA. In addition, we summarized the different benign comorbidities reported for reference by other studies ( Figure 2 ).

A total of 314 patients were described with symptoms, of whom 49.7% were asymptomatic or accidental findings, 39.2% experienced upper abdominal discomfort, and 5.4% developed fever, as shown in Table 2 . The spleen size was measured in 366 LCA patients, and 69.7% of them showed splenomegaly. Approximately 48.8% (127/260) of the patients had thrombocytopenia, while 31.4% (69/220) had anemia. Given its malignant potential, splenectomy with long-term follow up is the recommended treatment for LCA. Of the 401 patients, 81.0 and 19.0% underwent open and laparoscopic splenectomy, respectively. Although the accurate diagnosis rate was only 83.3% (10/12), image-guided biopsy was a relatively safe method when spleen lesions were difficult to diagnose.

Furthermore, 174 patients had a clear prognosis ( Table 2 ). A total of 159 (91.4%) patients survived without recurrence or metastasis after undergoing splenectomy. However, three patients developed recurrence or metastasis (11, 15, 16). In fact, one patient developed tumor recurrence in the accessory spleen 7 years after undergoing splenectomy, which showed no significant change during the 6-month follow-up (16). According to our statistics, four patients developed LCA both in the spleen and accessory spleen (16, 43, 89, 90). Therefore, it was quite reasonable to find a recurrence of LCA on the unresected accessory spleen. Strictly speaking, this should not be considered as a true relapse or metastasis, and we would rather think that a new LCA has formed in the accessory spleen. In the second patient, the co-occurrence of LCA along with LCHE was detected during splenectomy 8 years ago; the patient unfortunately died of abdominal multiple recurrence and metastasis (11). However, we think that it was the LCHE and not the LCA that recurred and metastasized. Actually, only the third case could be narrowly regarded as recurrence and metastasis, which presented as hepatomegaly with multiple metastatic tumors after splenectomy 10 years and was confirmed by open liver biopsy (15). Even so, the possibility of atypical LCHE or LCAS could not be ruled out.

Sixteen patients died during the follow-up period (1, 11, 12, 15, 39, 53, 58, 59, 91, 92). Among them, eight patients died from concurrent primary malignancies, namely, 2 gastric cancers, 2 malignant lymphomas, 1 multiple myeloma, 1 colon cancer, 1 pancreatic cancer, and 1 ovarian cancer. Four died of postoperative complications. Two patients died at 1.5 and 6 years, respectively, after splenectomy due to unknown causes. The other two patients who died were due to the recurrence and metastasis of the tumor to the liver. In fact, one of them died due to the recurrence and metastasis of LCHE (11), while the other died of hemorrhagic cerebral infarction 21 months after the diagnosis of intrahepatic metastatic LCA (15). Obviously, these causes of death had no direct correlation with LCA.

First, the ultrasound (US) features of 92 LCA patients are summarized in Table 3 . LCA manifested as hyperechoic ( Figure 3A ), hypoechoic, isoechoic, and heterogeneous echo in 43.5, 35.9, 9.8, and 10.9% of the patients, respectively. On the color Doppler imaging of 25 LCA patients, the color flow signals were 76% of hypovascular and 24% of hypervascular. There were relatively few data on the contrast-enhanced ultrasound (CEUS) findings. All 8 LCA patients who underwent CEUS showed enhancement in arterial phase. Of the 7 patients with portal phase description, 6 patients had enhanced features in CEUS. Only 3 patients had a description of delayed phase, and 2 patients presented enhancement. Second, 214 LCA patients underwent computed tomography (CT). On nonenhanced CT, 84.5% showed hypodense ( Figures 3B, D ), 13.9% isodense, and 1.6% showed hyperdense lesions. On enhanced CT, 125 LCA patients underwent arterial phase imaging, 58.4% presented enhancement. Of the 153 patients with portal phase imaging, 77.1% showed enhancement. Of the 96 patients with delayed phase imaging, 94.8% characterized by delayed enhancement ( Figures 3E, F ). Third, the magnetic resonance imaging (MRI) features of 66 LCA patients were summarized. On unenhanced T1-weighted images (T1WI), the lesions were depicted as hypointense masses in 75.9% of 58 patients. Moreover, 17.2 and 6.9% of patients showed isosignal intensity and high signal intensity, respectively. On unenhanced T2-weighted images (T2WI), the splenic lesions mostly manifested as a high signal intensity ( Figure 3G ) in 76.6% of 64 patients. However, they also showed hypointense, isointense, and mixed signal intense masses in 18.8, 3.1, and 1.6% of the patients, respectively. Of the 30 patients with DWI description, 90% showed hyperintense lesions, and only 10% showed hypointense lesions. On enhanced MRI, 43 LCA patients underwent arterial phase imaging, and 62.8% presented enhancement. Of the 45 LCA patients with portal phase imaging, 88.9% showed enhancement. Of the 30 patients with delayed phase imaging, 93.3% characterized by delayed enhancement ( Figures 3H, I ).

In addition, 19 LCA patients underwent radionuclide imaging ( Table 3 ). In nine patients who underwent positron emission tomography/CT examination, 7 who developed the splenic lesions had no abnormal uptake (11, 32, 33, 35, 36, 65, 93) ( Figure 3C ). However, 2 patients showed high uptake of 18F-fluorodeoxyglucose (11, 94). The other 10 patients underwent single-photon emission computed tomography, of whom 6 showed 99 mTc uptake (76, 80, 95–98), 2 showed Ga uptake (47, 89), and 2 had undescribed radionuclides (28, 99). Based on these results, only 30% of the patients showed a high uptake.

The size and weight of the spleen slightly to moderately increased ( Table 4 ). In the splenomegaly group, the median splenic volume and weight were 1,140 cm³ and 575 g, with the maximum volume and weight being 5,040 cm³ and 4,018 g, respectively. Meanwhile, the median splenic volume and weight were only 389 cm³ and 187 g, respectively, in the normal-sized spleens. Furthermore, LCA could appear as multifocal lesions or a solitary nodule with a solitary-to-multiple ratio of 0.31 in 349 patients. The average diameter of solitary LCA was 6.4 ± 4.0 cm, with the maximum diameter being 21 cm (31). The average diameter of solitary LCA was only 3.3 ± 2.6 cm in the multiple LCA group, with the maximum diameter being 18 cm (1).

A total of 18 patients described the natural history of LCA. Detailed description of the period of LCA formation was provided for ten patients (19, 26, 38, 42, 45, 65, 70, 76, 100, 101). The median time for LCA formation was 15 m, with the shortest and longest being 26 d and 4 y, respectively. Eleven patients with untreated LCA survived during follow-up, which had no obvious symptoms of discomfort (16, 19, 43, 65, 71, 102–107). The median of follow-up time of patients with untreated LCA was 24 m, with the shortest and longest being 0.5 m and 9 y. Although significant enlargement of untreated LCA was observed in three patients, no metastasis occurred. Among them (105–107), two patients developed an increase of the tumor size in the second year, from 0.5 to 3.4 cm and 4.3 cm respectively, and one showed an increase in the tumor size of 9 cm in the ninth year.

Histologically, LCA showed sinus like anastomosing channels with an irregular lumen ( Figure 4 ). These channels were either papillary or cystic and lined with tall, plump endothelial cells (1, 3). These endothelial cells exhibited hemophagocytosis and lacked features of nuclear atypia or mitotic activity, which was an important basis to distinguish them from other malignant tumor cells. The electron microscopic examination of LCA showed polygonal tumor cells surrounded by blood vessels (108). Notably, the vessels were only made up of a lining of medium electron density homogenous basilar membrane, which lacked smooth muscles. The tumor cells of LCA had an abundant cytoplasm, namely, mitochondria, a rough endoplasmic reticulum, and lysosomes. Some of them showed intermediate filaments and lipofuscin bodies.

The immunohistochemical staining for littoral cells usually reveals a dual differentiation pattern for both endothelial and histiocytic markers (109) ( Figure 4 , Table 5 ). Of the histiocytic markers, CD163, CD68, lysozyme, CD4, and CD11c showed very similar staining pattern with positivity rates of 100, 99.7, 99.1, 94.7, and 87.5%, respectively. However, factor XIIIa did not react in any cells. Of the vascular markers, all of them reacted positively with CD31, BMA120, UEA-1, FLI1, LYVE1, VEGFR2, claudin-5, and LMO2. A very similar staining pattern was observed with factor VIII and ERG, with positivity rates of 98.4 and 97.2%, respectively. As a tumor suppressor gene, Wilms tumor-1 (WT-1) was involved in development of tissues from the inner layer of intermediate mesoderm and played a role in regulating angiogenesis and proliferation of vascular smooth muscle cells (110). However, all 33 LCA patients showed no reactivity with Wilms tumor-1 (WT-1) (11, 12). Only 54.3% in the 232 LCA patients stained positively for CD34.

CD8 is typically associated with cytotoxic T cells. Its expression by littoral cells highlights the architectural framework of the red pulp in a dendritic pattern (109). Approximately 11.7% of 93 LCA patients stained positively for CD8 (108, 111). Moreover, 57.1% in the 69 patients stained positively for CD21, which is a type I transmembrane protein found on B cells, follicular dendritic cells, pharyngeal and cervical epithelial cells. In addition, S100 is a specific protein used as a marker of neurogenic tumors in pathological diagnosis. In 19 of 49 patients, the tumor cells of LCA exhibited positivity for S100 protein.

Of the lymphatic endothelial markers, D2-40 showed no reactivity in the cells of LCA (108). Similar staining pattern was also observed with epithelium marker CK. Only one of the 14 patients exhibited positivity for epithelium marker EMA (29). As a marker of mesenchymal tissue, the expression of vim indicates that the tissue originates from the mesenchymal rather than the epithelium. Approximately 65.3% of 72 LCA patients stained positively for vim. Notably, 50 patients stained positively for Ki67 with a median rate of 3%, with the minimum and maximum being 0 (112) and 23% (35), respectively. Finally, we proposed a new immunohistochemical phenotype summarized as CD31⁺/ERG⁺/FVIII antigen⁺/CD68⁺/CD163⁺/lysozyme⁺/CD8⁻/WT1⁻ to enhance the differentiation from other primary or secondary splenic tumors.

Based on the objective evidence, we herein proposed a process of diagnosis and treatment for LCA ( Figure 5 ). For primary splenic space occupying lesions, non-invasive clinical examinations should be performed initially. If the lesion is considered benign, a regular follow-up for 3 to 6 months is recommended. If the nature of the tumor could not be determined or the possibility of metastasis could not be ruled out due to a history of malignant tumor, an image-guided biopsy is suggested. Then, if the tumor is confirmed as LCA without other indications for splenectomy, regular follow-up is still necessary. If it is considered as a malignant lesion, an image-guided biopsy or a direct splenectomy should be performed. In addition, for splenic tumor patients with abdominal discomfort, pathological uncertainty, or continuous tumor enlargement, splenectomy is the preferred treatment. Basically, in order to improve the diagnosis and treatment of this rare disease, regular follow-up is recommended for all patients with definite LCA.