Orbital IgG4 Related Disease (IgG4RD): Incidence and Accompanying Histological Features Using the Latest Consensus Criteria

Purpose: Orbital Immunoglobulin 4 (IgG4) related disease (IgG4RD) is a fibro-inflammatory condition that mimics sclerosing orbital inflammatory disease (OID). The recently published IgG4RD consensus criteria included its defining histological features. Using the published criteria, this study aims to describe the frequency of orbital IgG4RD and its histological features in OID biopsies over a 1 year period.


Introduction
Immunoglobulin (Ig) G4 related disease (IgG4RD) is a multiorgan fibro-inflammatory condition that encompasses retroperitroneal fibrosis, autoimmune pancreatitis and orbital inflammatory disease (OID) [1][2][3]4]. It is characterized by a fibrosing mass-like lesion with dense lymphoplasmacytic infiltrate rich in IgG4 plasma cells on histology [1][2][3]5]. In the orbit, florid reactive lymphoid hyperplasia and sclerosing OID are common differential diagnoses which may mimic IgG4RD. The recently published IgG4RD consensus criteria included its defining histological features and preferred method for quantification of tissue IgG4 [5]. The three major histological features describe include (1) A dense lymphoplasmacytic infiltrate, (2) Storiform fibrosis, (3) Obliterative fibrosis and the other histological features include (A) Phlebitis without obliteration and (B) Increased eosinophilia [6]. In addition, the statement highlighted that the absolute count of IgG4 plasma cells was the preferred technique, and discouraged the use of "eyeballing" estimates of IgG4: IgG ratios advocating instead comparisons of the absolute counts of IgG4 and IgG cells via the use of photographic analysis or by microscopy. Prior to this, the lack of a diagnostic criteria had led to over-and under-diagnosis of this disease, especially in small biopsies. Using this criteria, our study aims to describe the frequency of orbital IgG4RD and the accompanying histological features in OID biopsies over a 1 year period.

Methods
Thirty-seven consecutive cases of orbital and lacrimal gland biopsies for OID at the Histopathology department, Singapore General Hospital (SGH) and Singapore National Eye Center (SNEC) from January 2012-January 2013 were prospectively reviewed using our orbital inflammatory protocol where all biopsies were subjected to Hemotoxylin and Eosin (H&E) and Masson Trichrome (MT) stains.

Quantification of fibrosis and inflammation
Percentage of fibrosis was calculated by the comparison of area of fibrosis highlighted by the MT stain with total area of tissue. The degree of inflammatory infiltrate was graded in similarly. Those with significant fibrosis and/ or inflammatory infiltrate (significant fibrosis or inflammation was defined as >25% of the biopsy section) were subjected to immunohistochemical (IHC) analysis for CD138 positive plasma cells, tissue IgG and IgG4 in addition to routine diagnostic stains to exclude infections and lymphoma when warranted. The pattern of fibrosis and inflammation and the presence of vasculitis were recorded.

Inflammation
The predominant inflammatory infiltrate comprised of lymphocytes and plasma cells ( Figures 1C, 2C and 3B) and was present in about 6/10 (60%) cases. Thirty percent showed only lymphoplasmacytic infiltrate without significant fibrosis (<25% biopsy area), another 30% showed BOTH dense lymphoplasmacytic infiltrate with significant storiform fibrosis (>50%) and the remaining 40% did not have significant inflammation but only storiform fibrosis as the main feature. The predominance of plasma cells was often only highlighted on CD138 IHC ( Figures 1D, 2D and 3D) and was not always apparent on routine HE stain. Scattered eosinophils were noted in the background but was only significantly elevated in 1 case ( Figure 3B). Xanthogranulomatous inflammation in the absence of other infectious or underlying etiology was also noted in another case ( Figure 2B and Table 1).

Vasculitis
Active vasculitis and obliterative phlebitis was not seen in any of the cases. Perivenular fibrosis and perivascular "onion skinning" was present in all 10 cases suggestive of previous non-obliterative phlebitis. (Table 1)

Tissue IgG4 plasma cell counts and IgG4: IgG ratio
In the 3 cases with BOTH storiform fibrosis and dense

Quantification of tissue IgG4 and calculation of IgG4: IgG ratio
Using consecutive sections, the numbers of absolute counts of CD138 positive IgG4 and IgG cells were counted in 3 separate high power fields (x40) where inflammation was maximal. The mean number of IgG4 and IgG was then calculated, and the ratio of IgG4:IgG was determined based on the mean number of IgG4 and IgG. Exclusion criteria included the diagnoses of Granulomatosis with polyangitis (GPA, previously known as Wegener's granulomatosis), Sjogren's syndrome, adnexal lymphomas such as extranodal marginal lympoma and cases with neutrophilic infiltration, necrosis or granulomatous inflammation. The histological diagnosis of IgG4RD was highly suggestive when the at least 1 of the major histological feature of the consensus criteria (storifiorm fibrosis, dense lymphoplasmacytic inflammation and obliterative phlebitis) was present with elevated tissue IgG4 plasma cells as defined by the consensus criteria [6]. The diagnostic criteria for elevated tissue IgG4 included both the presence of: 1) CD138 positive IgG4 positive plasma cells >50 per high power field (hpf), and 2) Ratio of IgG4: IgG >40% as stated in the consensus criteria [6]. These confirmed cases were also reviewed clinically and radiologically for evidence of systemic IgG4RD.

Statistical analysis
All data were expressed as mean ± SD.

Results
Ten of 37 (27%) of our orbital biopsies showed features suggestive of sclerosing orbital inflammation or IgG4RD characterized by significant fibrosis and/ or lymphoplasmacytic inflammation. The remaining cases showed only non-specific chronic inflammation or reactive lymphoid hyperplasia. Using the histological consensus criteria, only 3 of these (3/10, 30%) cases fulfilled the morphological features and IgG4 IHC definition described above. The histological patterns included: 1) Sclerosing orbital inflammation or dacroadenitis (Figures 1A-1C ratio of 10-20% (Table 1).

Systemic review
Our 3 cases of histologically diagnosed as IgG4RD had no prior history of IgG4RD, but were subsequently all found to have elevated serum IgG4 and other IgG4RD features: involvement of the parotid gland (Case 1, sclerosing dacroadenitis), autoimmune pancreatitis (Case 2, sclerosing xanthogranulomatous orbital inflammation) and retroperitoneal fibrosis (Case 3, EAF). lymphoplasmacytic inflammation, a marked increased of tissue IgG4 plasma cells that fulfilled both criteria with the average of 119 (range: 102 to 134 cells) IgG4 positive plasma cells (identified using CD138, plasma cell marker)/ hpf and an IgG4: IgG ratio of 75-90% (Table 1).
In the remaining cases where only either storiform fibrosis (4 of 7) OR dense inflammation (3 of 7) was present, the IgG4 plasma cell count and IgG4: IgG ratio failed to meet the criteria with the average of 12 (range: 4 to 25 cells) IgG4 positive plasma cells/ hpf and an IgG4: IgG  with elevated IgG4 plasma cells however does raise the issue of whether another xanthogranulomatous orbital inflammation, Adult onset xanthogranulomatous inflammation could represent IgG4RD as both share the association with asthma and elevated IgE, although more cases would have to be evaluated to confirm this observation.
In our study, the most common major histological feature seen for IgG4RD was storiform fibrosis 70%, followed by dense lymphoplasmacytic infiltrate (60%). The lack of storiform fibrosis in the lacrimal gland as described by the consensus group [6] was not reflected in our series as 70% of our cases presented with significant fibrosis. This discrepancy may represent the different stages of orbital inflammation at the time of biopsy, however, a larger series would be necessary to confirm this.
When both storiform fibrosis and lymphoplasmacytic infiltration were present, we noted a marked elevation of IgG4 plasma cells and the IgG4: IgG ratio which was consistently almost 2-fold above the diagnostic criteria with over 100 IgG4positive plasma cells present and the IgG4: IgG ratio of 75-90% (Table 1). This correlates with the consensus group findings and others who report the number of IgG4 plasma cells in the lacrimal gland as typically in the 100s range (Table 1) [6,10]. This should be kept in mind when only a borderline number of IgG4 plasma cells are present in the biopsies to prevent over reporting. In such situation, it is necessary to exclude other autoimmune (AI) conditions since IgG4 plasma cells are not specific to IgG4RD and to correlate with the presence of other systemic IgG4RD features.
Although borderline values of IgG4 plasma cells may represent the "early stage" of IgG4RD, long term follow up to determine the true IgG4 nature in repeat or other organ biopsies would be necessary to confirm such a suspicion in the absence of other clinical features.

Conclusions
Rootman et al. first described an idiopathic sclerosing orbital inflammation which is clinically aggressive [7]. Orbital IgG4RD has renewed interest in sclerosing orbital inflammation and reports have suggested their shared histological features may represent a spectrum of IgG4RD [4,8,9].
Using the IgG4RD consensus criteria, we found that almost 27% of our OID orbital biopsies shared overlapping features of sclerosing and IgG4RD making this criteria useful for differentiation of these entities since sclerosing OID is a more aggressive orbital inflammation requiring early immunosuppression therapy [7] in comparison to IgG4RD which is highly steroid sensitive, in addition to having systemic implications [1,4,5,8]. In our study, IgG4RD accounted for 11% of all our OID and up to 30% of our orbital inflammation with significant fibrosis and/ or inflammation. As IgG4RD overlaps with many other systemic AI disorders, it is a diagnosis of exclusion. Elevated tissue IgG4 is also present in Sjogren's syndrome and GPA related orbital inflammation and must be excluded prior the diagnosis of IgG4RD as in our series. [4,6] Of note, in our study, the orbital biopsy provided the first histological evidence of IgG4RD and it is important for the ophthalmologist to exclude systemic involvement.
The histological presentation of IgG4RD is variable from sclerosing dacroadenitis, to defined entities such as EAF, which recently has been suggested as IgG4RD [9,11,12]. Xanthogranulomatous inflammation has not been previously described in IgG4RD and extensive workup to exclude infection and other autoimmune disorders is necessary. Our patient with sclerosing xanthogranulomatous IgG4RD was subsequently found to have autoimmune pancreatitis and elevated serum IgG4RD. The association of xanthogranulomatous inflammation Cases Field 1 (40x) Field 2 (40x) Field 3 (40x)

Mean ± standard deviation (SD)/ 3hpf
Of the cases that did not meet the diagnostic criteria of IgG4RD, four cases had only significant fibrosis whilst 3 cases had predominantly inflammation without significant fibrosis and all lacked tissue IgG4 plasma cells (an average of 12 IgG4 plasma cells/hpf and an IgG4: IgG ratio <25%, ( Table 1). The issue of whether the 4 cases with only significant "sclerosing" inflammation could represents a "burntout" fibrotic stage of IgG4RD rather than "idiopathic sclerosing inflammation" has yet to be resolved. However, without serum IgG4 elevation and other clinical evidence of IgG4 RD, it may be best to classify these cases as "sclerosing orbital inflammation with insufficient evidence for IgG4RD".
In conclusion, orbital IgG4RD represents 11% of all our OID and 30% of those with significant sclerosis and dense lymphoplasmacytic infiltrate seen over 1 year. Tissue IgG4 levels is paramount to the diagnosis and both the number of IgG4 plasma cells and IgG4: IgG ratio must be assessed. Systemic review is a useful adjunct to exclude IgG4RD. Although storiform fibrosis and lymphplasmacytic inflammation is predominantly seen, the presence of eosinophils and xanthogranulomatous inflammation may also be present in orbital IgGRD.