Characteristics of cranial vault lymphoma from a systematic review of the literature.

Background: Cranial vault lymphomas are rare and their clinical features are often similar to those of cranial vault meningiomas. The objective of this review was to identify the features helpful for differentiating lymphomas of the cranial vault, from meningiomas which were the most common diagnosis before the definitive pathological diagnosis.
Methods: The inclusion criterion was a histologically proven malignant lymphoma initially appearing in the calvarium. We conducted a literature search of the electronic PubMed and Ichushi-Web databases up to June 1, 2020. Cranial vault lymphoma that was diagnosed after an original diagnosis of lymphoma in a nodal or soft-tissue site was excluded from the study. Descriptive analyses were used to present the patient characteristics.
Results: A total of 111 patients were found in 98 eligible articles. Almost all studies were case reports. The most common symptom was a growing subcutaneous scalp mass (84%) present for a mean duration of 5.9 months before the patient presented for treatment in analyzable cases; this fast growth may distinguish lymphomas from meningiomas. The tumor vascularization was often inconspicuous or poor, unlike well-vascularized meningiomas. A disproportionately small amount of skull destruction compared with the soft-tissue mass was observed in two-thirds of the analyzable cases.
Conclusion: This qualitative systematic review identified several features of cranial vault lymphomas that may be useful in differentiating them from meningiomas, including a rapidly growing subcutaneous scalp mass, poor vascularization, and limited skull destruction relative to the size of the soft-tissue mass. Copyright:

INTRODUCTION

Malignant lymphoma of the bone is uncommon and, hence, presents diagnostic and therapeutic problems.[71] Cranial vault involvement has been reported to account for 5.5% of bone lymphoma cases and the number of reported cases of lymphoma initially appearing in the calvarium is limited.[71] The diagnosis of this rare cranial vault tumor before pathological diagnosis has usually been not lymphoma but meningioma, metastatic tumor, or other mesenchymal tumors, which has tended to lead to radical operative treatments like gross total removal of the tumor.[1,25,96] If we can make a differential diagnosis of cranial vault lymphoma preoperatively, we may choose less radical procedures like biopsy and can start chemotherapy at an earlier time. This article presents a systematic review of the literature to show demographic, clinical, and imaging characteristics, as well as treatments and outcomes of cranial vault lymphoma. We specify characteristic features of cranial vault lymphoma that may be helpful for their diagnosis, especially to distinguish them from meningiomas, which have been the most common diagnosis of cranial vault lymphomas before definitive pathological diagnosis.[1,2,4,13,35,36,37,50,54,65,66,70,75,84,86,88,96,105]

MATERIALS AND METHODS

Eligibility criteria

The inclusion criterion was a histologically proven malignant lymphoma initially appearing in the calvarium. Cranial vault lymphoma that was diagnosed after an original diagnosis of lymphoma in a nodal or soft-tissue site was excluded from the study. Skull base lymphoma and dural lymphoma were also excluded from the study. Articles whose full text was unable to be located were excluded from the study. We excluded systematic and retrospective review articles and case series articles that did not include case-specific data.

Information sources and search strategy

We conducted a literature search of the electronic PubMed and Ichushi-Web databases up to June 1, 2020, using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. We extracted all human reports on cranial vault lymphomas, using the following terms: cranial vault lymphoma; skull lymphoma; calvarial lymphoma; transcalvarial lymphoma; dural lymphoma; and combinations of the variables of lymphoma, lymphosarcoma (an obsolete classification of non-Hodgkin lymphoma), reticulum cell sarcoma (another obsolete classification of non-Hodgkin lymphoma), cranial, vault, skull, and calvarium. The search resulted in 1492 PubMed citations and 391 Ichushi-Web citations, 113 of which were duplicated, resulting in a total of 1770 articles [Figure 1].

Figure 1:

The PRISMA flowchart of our systematic review. New studies were those identified from the reference lists of articles identified in the initial screen.

Selection process

Two reviewers (N.N. and T.F.) independently and in duplicate screened, reviewed, and discussed all the selected articles. The full text of all eligible articles was reviewed, and their data extracted and collated. In cases where questions regarding the inclusion of certain articles arose, this was discussed with a third reviewer, K.N.

Data collection process and data items

Data of the eligible works were obtained through careful analysis of the full text by two authors (N.N. and T.F) independently. The senior author, K.N., was available in case of a split decision. Questions arising as to pathological diagnosis were discussed with a pathologist, S.M. We analyzed the clinical and radiological characteristics of the patients, as well as their treatments and survival in these published studies. Specifically, we extracted the following items: age of the patient; sex; clinical symptoms; location of tumors; findings of palpation; skin condition; speed of growth; existence of other lesions; types of treatment; extent of resection; duration of follow-up and outcomes; laboratory data; imaging data of skull X-ray, angiogram, computed tomography (CT), magnetic resonance imaging (MRI), and others; bone images on CT; extra- and intracranial tumor extension on CT and MRI; dural tail and brain invasion on MRI; and histopathological types. Due to the heterogeneity of patient descriptions, some clinical and imaging features were not explicitly reported for each patient. We extracted and reported only unambiguously described data. Data on clinical and imaging features were also extracted from the patient imaging data. Evaluation of publication bias was not feasible because of heterogeneity and because most of the included studies were case reports and case series.

Statistical analysis

Descriptive analyses were used to present the patient characteristics. Continuous variables were expressed as mean ± standard deviation, and categorical variables were expressed as number and percentage. Because of the limited follow-up data included in each intervention and the lack of standardization of assays and treatments across the many laboratories included in the present review, we were unable to statistically compare the findings on images and the changes in clinical outcomes. All calculations were performed with JMP 13.2.1 (SAS Institute, Inc., Cary, North Carolina, U.S.).

RESULTS

From among the articles found, we selected all studies reporting patients with cranial vault lymphoma (n = 106) without limitation of language and identified additional studies from the reference lists of the articles (n = 3). After discarding duplicate references and publications (n = 2), as well as excluding reports of secondary cranial vault lymphoma, skull base lymphoma, or dural lymphoma (n = 9), we settled on 98 articles for careful review. The number of articles retained at each stage of data acquisition is shown in a PRISMA flowchart [Figure 1].

We found 111 cases of histologically proven malignant lymphomas initially appearing in the cranial vault in the 98 articles and analyzed the data of 111 patients [Tables 1, S1, and S2].[1-11,13-16,19-25,27-56,58-70,72-76,78,80,83-89,91-109]

Table 1:

Clinical features, treatment, and outcome of 111 patients with cranial vault lymphoma.

The average patient age was 52 ± 20 years (range, 3–85 years) [Table 1]. The male-to-female ratio was 1.09:1. Common symptoms were a growing subcutaneous mass on the scalp (84%), headache (33%), focal neurological deficit (25%), and seizure (6%) in the 110 cases with available data. The parietal (57%), frontal (54%), occipital (21%), and temporal (15%) bones were the affected sites in 110 cases with analyzable data. The subcutaneous scalp mass was firm (57%), soft (16%), nontender (56%), tender (22%), or accompanied by reddish skin or ulcer (7%) in 68 cases with a description of the scalp mass. In the 49 cases with a description of mass growth, the mean duration of growth was 5.9 ± 6.3 months before the patient presented for treatment. In 23% of 93 cases with available data, lesions outside the cranial vault were present.

Laboratory data were described in 59 cases, with human immunodeficiency virus positivity in 5 cases (8%), elevated lactate dehydrogenase in 6 cases (10%), anemia in 4 cases (7%), increased white blood cell count in 3 cases (5%), elevated erythrocyte sedimentation rate in 2 cases (3%), elevated alkaline phosphatase in 1 case (2%), increased soluble interleukin-2 receptor (sIL2R) in 1 case (2%), and decreased platelets in 1 case (2%) [Table S1]. Findings on plain skull X-ray were described in 38 cases, including osteolytic changes at the tumor site in 29 cases (76%), no change in 6 cases (16%), hyperostosis in 2 cases (5%), and periosteal reaction in 4 cases (11%). A sharp margin of the skull lesion was observed in 6 cases (16%) and an indistinct or irregular margin of the skull lesion in 15 cases (40%). In 12 cases with description of tumor vascularization on angiogram, no or poor tumor vascularization was found in 7 cases (58%). If any vascularization was present, it was mainly derived from the external carotid artery circulation. Among 32 cases with analyzable CT data, the tumor was hyperdense in 25 cases (78%), isodense in 5 cases (16%), and hypodense in 2 cases (6%). Among 42 cases with contrast-enhanced (CE) CT data, the tumor was enhanced well in 37 cases (88%) and slightly in 5 cases (12%). Most tumors were enhanced diffusely, either homogeneously or heterogeneously, whereas some extracranial components showed peripheral enhancement. On MRI, T1-weighted imaging (T1WI) with analyzable data (n = 38) showed a hyper- and isointense tumor in 1 case (3%), isointense tumor in 11 cases (29%), iso- and hypointense tumor in 4 cases (11%), and hypointense tumor in 22 cases (58%). T2-weighted imaging (T2WI) (n = 40) showed a hyperintense tumor in 21 cases (53%), hyper- and isointense tumor in 1 case (3%), isointense tumor in 12 cases (30%), iso- and hypointense tumor in 5 cases (13%), and hypointense tumor in 1 case (3%). CE-T1WI (n = 48) showed an enhancing tumor in 48 cases (100%). The tumors tended to show uptake on 18F-fluorodeoxyglucose positron emission tomography (FDG-PET) and on bone scintigraphy and gallium scintigraphy, which were also used for the evaluation of lesions outside the cranial vault.

Extra- and intracranial extensions of the tumor were evaluated by CT, MRI, or both in 87 cases [Table S2]; the extracranial extension was much larger than the intracranial extension in 35 cases (40%), both were nearly the same size in 35 cases (40%), the intracranial extension was much larger than the extracranial extension in 16 cases (18%), and there was neither intra- nor extracranial extension in 1 case (1%). Bone changes of the cranial vault due to the tumor were evaluated by CT, MRI, or both in 84 cases. The images showed no skull changes and/or preserved skull contour in 11 cases (13%), osteolytic changes in 68 cases (81%), hyperostosis in 4 cases (5%), and only sclerosis at the lesion in 1 case (1%). In the 68 cases with osteolytic skull changes due to the tumor, the skull was penetrated or dissolved to less than half of the thickness in 26 cases (38%), whereas permeative dissolution with relatively preserved skull contour was observed in 35 cases (51%). Periosteal bone formation was observed in 10 cases (15%). In seven cases, there was no detailed description. A disproportionately small area of cortical destruction of the cranial vault relative to the volume of the extra- or intracranial soft-tissue mass, which we defined as cortical destruction less than one-fifth of the soft-tissue mass in diameter, was observed in 50 cases (67%) on CT, MRI, or both (n = 75). On MRI (n = 57), a dural tail was observed in 42 cases (74%), and invasion of the brain was observed in 15 cases (26%).

Surgery was detailed in 98 cases, of which 45 (46%) involved a biopsy or partial removal and 53 (54%) involved subtotal or gross total removal [Table 1]. Adjuvant therapy was reported for 96 cases. The breakdown was radiotherapy alone for 24 cases (25%), chemotherapy alone for 34 cases (35%), and both for 38 cases (40%).

The type of lymphocyte was described in 80 cases: 75 (94%) were B-cell lymphomas and 5 (6%) were T-cell lymphomas [Table S2]. Diffuse large B-cell lymphoma (DLBCL) was the most common, being reported in 34 cases.

To determine the rates of survival, we excluded cases in which the patients died of a cause unrelated to the skull lesion and included cases with follow-up periods of more than 6 months (n = 74) or 1 year (n = 54). In the included cases, 67 (90.5%) and 45 (83.3%) patients were alive at the follow-up of 6 months and 1 year, respectively [Table 1].

DISCUSSION

We reviewed the demographic, clinical, and imaging characteristics of cranial vault lymphoma to specify the features that might be helpful for differential diagnosis, especially between lymphomas and meningiomas of the cranial vault.

When a tumor with intra- and extracranial extension sandwiching the skull is seen, meningioma with extracranial extension is often first suspected. In many cases, the subcutaneous scalp mass was firm and nontender, which is also similar to meningioma with extracranial extension. However, in our review, the subcutaneous scalp mass grew very rapidly before the patient presented for treatment, for a mean duration of 5.9 months, which is atypical for meningioma, which is generally slow growing.[18,79]

Laboratory data were unremarkable in many cases. Kosugi et al.[48] reported elevated sIL2R, a marker of lymphoproliferative neoplasms,[81] in a patient with cranial vault lymphoma, suggesting that elevated sIL2R associated with a cranial vault tumor might indicate lymphoma rather than meningioma. Another finding that could be used to differentiate the two tumor types is poor or no tumor vascularization on angiograms, which differs from the rich tumor vascularization from the external carotid artery circulation observed in many cases of meningioma.[90] The tumor also showed high uptake on FDG-PET.[4,11,40,86,92,96] Because meningiomas are mostly slow-growing tumors and their glucose metabolism might be only moderately elevated, the high physiological glucose uptake of the normal cerebral cortex leads to a low meningioma-to-background ratio.[26] Hence, a high tumor-to-background ratio may indicate a lymphoma and contribute to ruling out a meningioma.

The extracranial component tended to be at least as large as the intracranial component in cranial vault lymphoma. Because meningioma usually originates from the meninges, theoretically, it tends to grow intracranially rather than extracranially. Extracranial-dominant extension might, therefore, also contribute to distinguishing cranial vault lymphoma from meningioma.

However, other findings are unlikely to clarify the diagnosis. For example, the tumors tended to be hyper- to isodense on CT, iso- to hypointense on T1WI, and hyper-to isointense on T2WI, which are nonspecific features of skull tumors. Although a dural tail was observed in many cases, as reported by Xing et al.,[108] this finding is also observed in meningiomas.[79]

It has been suggested that lymphoma cells infiltrate the spaces within the diploe and extend along the emissary veins to infiltrate the soft tissues on either side of the bone.[5,20,30,77] Osteolytic changes were most frequently observed on CT and MRI (68 cases) followed by no changes and/or preserved skull contour (11 cases). In the cases with osteolytic skull changes, permeative dissolution with relatively preserved skull contour was observed in 35 cases. Although the destruction of the cranial vault tended to be small, the extracranial and intracranial components of the tumor tended to be large. These data support the characteristic permeating growth pattern of cranial vault lymphoma involving a large soft-tissue component and slight bony destruction, as reported previously.[20,62,90,108]

After histological confirmation of the diagnosis, patients were usually treated with adjuvant chemotherapy, radiotherapy, or both. Although the best treatment for cranial vault lymphoma has not been elucidated because of the paucity of cases and lack of clinical studies, treatments based on those for systemic malignant lymphoma tended to be adopted.

B-cell lymphomas accounted for approximately 94% of the cranial vault lymphomas, whereas T-cell lymphomas roughly accounted for the remaining 6% of cases. Peripheral T-cell lymphomas account for 6–10% of all cases of non-Hodgkin lymphoma[57] and T-cell lymphomas account for 5–6% of cases of primary lymphoma of bone.[12,82] However, T-cell lymphomas account for only 2% of all primary lymphomas in the central nervous system (CNS),[17] suggesting that cranial vault lymphomas are more similar to extracranial lymphomas than they are to CNS lymphomas. We could not determine the subtypes of lymphoma of all cases using the present classification because our study included some very old reports from before the discovery of lymphocyte markers. Nonetheless, DLBCL accounted for the majority of cases of cranial vault lymphoma in the available data, which is also true for cases of bone lymphoma.[82]

In analyzable follow-up data, 67 and 45 patients were alive after 6 months (n = 74) and 1 year (n = 54), respectively [Table 1]. Although our review does not allow for predicting prognosis due to the limited numbers and the limited duration of follow-up data, the 6-month survival rate and 1-year survival rate are not <60.4% and 40.5%, respectively.

In the present review, we have provided objective data from patients reported in published studies. The difficulty in systematically reviewing the data reported in the literature is the heterogeneity of the data. In particular, almost all reports reviewed for this study were case reports. Because of the limited number and the heterogeneity of described data of preoperative findings and treatments, we could perform only descriptive analyses. Another weakness of the present review resulted from biased reporting in the published studies, with likely underreporting of recent cases in the United States and European countries because of decreased novelty in inverse proportion to the accumulation of reported cases. The data that were more likely to be reported, if present, included findings on MRI and CT. Findings of plain skull X-ray and angiography and long-term follow-up data were less likely to be reported.

The strength of the present review is the comprehensive nature of the literature search. We analyzed cases from all over the world, reported not only in English but also in four non-English languages.

CONCLUSION

Cranial vault lymphoma is a rare entity among skull tumors. To the best of our knowledge, this is the largest reported pooled database describing cranial vault lymphoma patients. The most common symptom was a rapidly growing subcutaneous scalp mass. The tumor was poorly vascularized on angiography. Skull destruction on images was mild and disproportionately small despite the large size of the extracranial and/or intracranial component in two-thirds of the cases. These features should help to distinguish lymphoma from meningioma.

Table S1:

Laboratory data and features on images of patients

Table S2:

Bone imaging, dural tail, brain invasion, and pathological features of patients with cranial vault lymphoma.