Localized immunoglobulin light chain amyloidosis: Novel insights including prognostic factors for local progression

In localized light chain amyloidosis (locAL), amyloidogenic light chains (aLC) are produced and deposited locally by a B‐cell clone. We present 293 patients with immunohistochemically confirmed locAL. Lung (nodular pulmonary) with 63 patients was the most involved organ. The aLC was λ in 217 cases (κ:λ ratio 1:3). A local B‐cell clone was identified in 30% of cases. Sixty‐one (21%) had a concomitant autoimmune disorder (cAD). A monoclonal component (MC) were present in 101 (34%) patients and were more frequent in subjects with cAD (51% vs 34%; P = .03). Cigarette smoking was more prevalent in lung locAL (54% vs 37%; P = .018). After a median follow‐up of 44 months, 16 patients died and 5‐ and 10‐years locAL progression‐free survival (PFS) were 62% and 44%. Interestingly, locAL‐PFS was shorter among patients with an identified clonal infiltrate at amyloid deposition site (40 vs 109 months; P = .02) and multinuclear giant cells and/or an inflammatory infiltrate resulted in longer locAL‐PFS in lung involvement (65 vs 42 months; P = .01). However, no differences in locAL PFS were observed in patients with cAD, a MC and involved organ site. Treatment was administered in 163 (54%) patients and was surgical in 135 (46%). Median locAL‐PFS after first treatment was 56 months. Responders had longer locAL‐PFS (78 vs 17 months; P < .001). Three patients with lung locAL and a MC were diagnosed as systemic AL amyloidosis at follow‐up. In summary, locAL pathogenesis seems to be heterogeneous and the clonal infiltrate leads local progression.


| INTRODUCTION
Amyloidoses are a group of rare diseases caused protein misfolding and deposition in organs and tissues as amyloid fibrils. Thirty-six different amyloidogenic proteins have been identified to date in human. 1 In systemic amyloidosis, the amyloidogenic precursor is a plasma protein that might target multiple organs. 2 The most frequent type of systemic amyloidosis in the Western countries is systemic immunoglobulin light chain (AL) amyloidosis (sysAL). 3 This disease is caused by a light chain (LC) produced by a B-cell clone in the bone marrow. 4 However, immunoglobulin LCs are also responsible of local AL amyloidosis (locAL). This is a very rare and less studied disease, accounting for approximately 10% of all amyloidosis cases in referral centers (no other population data available).
In locAL, locally produced LCs deposit at a single anatomic site, forming one or multiple tumor-like amyloid lesions (amyloidomas). 5,6 Our knowledge on locAL has improved in the past 5 years thanks to description of two large case series. The National Amyloidosis Center (NAC) described for the first time the natural history of this disease in 606 patients. 7 Bladder, larynx and skin emerged as the most commonly involved organs. Furthermore, patients with locAL may also present concomitant autoimmune disorders (cAD), especially Sjögren syndrome (SjöSy), lymphoproliferative diseases, mainly marginal zone lymphoma (MZL), and a monoclonal gammopathy of undetermined significance (MGUS). [8][9][10][11] Recently, a report from the Mayo Clinic's group added valuable information about response to therapy and local progression from 413 cases of locAL. 12 Although life expectancy in locAL was comparable to the general population, the clinical history was characterized by frequent local progressions requiring further treatment. Importantly, repeated surgical interventions were often cause of consistent morbidity and quality of life impairment.
Currently, it is not clear if some peculiar characteristics of this complex and heterogeneous disease may affect the prognosis in locAL, particularly local progression. Moreover, few data are available on local cellular infiltrate and B-cell clone at amyloid deposition site and its role in natural history of locAL has not been studied so far. 13 We present the results of a comprehensive study conducted on a large series of 293 consecutive patients with verified locAL evaluated at the Heidelberg Amyloidosis Center. We coupled an extensive characterization of clinical features with detailed pathology data from tissue biopsy. Finally, we described treatment and outcome and studied local and systemic factors that may affect local progression for the first time.

| Patients and inclusion criteria
Study design and patient population are summarized in Figure S1. The prospectively maintained database of the Heidelberg Amyloidosis Center was searched for patients with locAL evaluated between 04/ 2000-10/2019. Only patients with conclusive immunohistochemical typing and clinical findings for locAL were included in the study.
A multifocal involvement was defined as the presence of multiple amyloid lesions in the absence of systemic involvement. SysAL was ruled out in all patients with evaluation of clonal and organ biomarkers. In selected cases, echocardiogram and abdominal fat pad aspirate were performed (see Supplement).
All patients gave written informed consent for their clinical data to be used in retrospective studies in accordance with the Declaration of Helsinki.

| Response to treatment and local progression
Follow-up of amyloidomas was performed by clinical, radiographic and endoscopic examination. Response to treatment and local progression were defined according to changes in symptoms and/or size of the amyloidomas. Particularly, changes in size of the amyloidomas at imaging were evaluated for response and progression in asymptomatic patients. Modifications of imaging and endoscopic findings over time was assessed and documented by local radiologists or physicians.
Local progression was evaluated in patients from diagnosis and after first treatment. This was done in order to better evaluate the natural history of locAL and the impact of different therapeutic strategies and response to treatment. Progression to sysAL was defined by detection of the amyloidogenic light chain (aLC) in the serum and/or urine and onset of another involved organ site with detection of the aLC in abdominal fat pad or organ biopsy.

| Amyloid typing and cellular infiltrate characterization
A sample of tissue biopsy or a surgical resection from the involved organ site were sent to a Pathology Unit for amyloid identification and typing. Two-hundred forty-six (84%) samples were sent to the Department of Pathology at the University Hospital Kiel, that has proven its expertise on immunohistochemical typing of amyloid and identification of amyloidogenic clones on tissue biopsies with novel sensitive techniques. In 180 cases, abdominal fat pad aspirate was also obtained and evaluated. Immunohistochemical typing of amyloid was performed on tissue biopsy with custom made antibodies in all patients, as previously described. 14 Reports from the Pathology Unit were systematically reviewed for data collection. Data on biopsy size were collected from 189 patients and information about cellular infiltrate at amyloid deposition site was available in 154 cases. B-cell clonality was assessed by immunohistochemistry, in situ-hybridization or PCR-based immunoglobulin heavy chain (IGH) and light chain (IGK) gene rearrangement analyses, as described. 15

| Statistical analyses
Overall survival (OS) and local progression free survival (locAL-PFS) curves were plotted according to Kaplan Meier, and differences in survival 3 | RESULT

| Patients characteristics
Patients characteristics are reported in Table 1 and detailed information on organ involvement are summarized in Table 2. Lung (nodular pulmonary) was the most commonly involved organ. Patients with lung locAL were older at diagnosis (68 vs 55-year-old; P < .001), had more frequently multifocal involvement (62% vs 40%; P = .026) and presented a high prevalence of smokers (54% vs 37%; P = .014).
A female preponderance was observed among patients with skin, soft tissues (ST), eye and central nervous system (CNS) locAL.
Lung involvement was particularly frequent in 65 asymptomatic patients (72%) in which locAL was diagnosed incidentally during regular medical evaluation for a concomitant comorbidity or, in a few cases, during the radiological assessment of an acute trauma.

| Identification of the amyloidogenic light chain and cellular infiltrate
Median length of the histological material from tissue biopsy or surgical resection was 1.2 cm (range 0.1-14.5 cm). Larger samples were obtained from patients with lung locAL (median length 2.5 cm vs 1 cm; P < .001), while smaller specimens came from patients with lower airways locAL (median length 0.6 cm vs 1.5 cm; P = .001). λ aLC was identified by amyloid typing on tissue in most cases with a κ:λ ratio of 1:3. Interestingly, we observed an organ site specific variation of κ:λ ratio (

| Response to treatment
Details on treatment and outcome according to organ involvement are summarized in Table 3 and Figure S1C. The treatment approach was decided by local physicians in most cases. A "watch and wait" strategy was adopted in 130 cases. One hundred sixty-three (56%) patients received a treatment, which was surgical in 135 (83%). Treated patients were younger (62 vs 67-year-old; P = .013), with more urinary tract (17% vs 8%; P = .034) and less GI (5% vs 21%; P < .001) and lower airways locAL (8% vs 15%; P < .001) (see Table S1).

| Overall survival from diagnosis
After a median follow-up of 44 months, 16 patients died (9 with lung, 4 with lower airways and 3 with nasopharynx locAL). Ten of these subjects were also smokers. Median OS at 5-and 10-years was 94% and 92% ( Figure 1A). OS was poorer in patients with lung locAL (5-years OS 79% vs 97%; P < .001). Death was attributed to locAL in only one case.
Not-amyloid related causes of death were neoplasia (2 patients In 9 (3%) patients, multiple amyloid deposits extended through more than one organ. In 7 cases, amyloid was disseminated through the respiratory tract. In one case locAL presented as nodular amyloidomas in the skin (skin and ST involvement) and in the oral mucosa (GI involvement). In another patient, amyloid deposits were found in nasopharynx (respiratory involvement) and in a lateral cervical lymph node (lymphatic tissue involvement). This technique has been proven adequate and effective for amyloid typing and was validated in patients with sysAL. 14 As a further confirm, in all patients the LC of the local clone was the same as aLC. The second peculiar aspect of our study is the histological and molecular characterization of the cellular infiltrate at amyloid deposition site.
Our findings confirm and extend data from previous studies.
Patients with lung locAL are usually older, 12 and skin and ST locAL are more common in women. 16 A cAD and a MC and/or abFLCR are particularly common in these patients. In general, prognosis of locAL is good, however, a worse survival is observed in patients with lung locAL. However, considering also that the majority was asymptomatic at diagnosis, it seems more likely that in these cases death was related to their older age or other comorbidities. Response to treatment and local progression were evaluated according to changes in clinical manifestation, imaging and endoscopic findings, as reported before by the Mayo Clinic group. 12 Most patients responded to treatment but 40% showed local progression. Five-years locAL-PFS was similar among patients with different involved organ sites. As expected, in treated patients a response resulted in longer locAL-PFS. This was confirmed both when response was assessed with clinical and radiological evaluation (ie in symptomatic patients) or with imaging alone (ie in asymptomatic patients). At 5 years from treatment, 41% of responders and 71% with stable disease had a local progression. We confirm the effectiveness of radiotherapy, particularly in combination with surgical debulking, in selected cases. [17][18][19] However, in general we observed no differences in outcome according to different treatment regimens.
Progression to sysAL seems to be a rare event. Similar to NAC's col- involvement, which can be affected in both, locAL and sysAL. [20][21][22] One difference to past studies relate to the frequency of organ involvement. In our series, the lung is the most commonly involved anatomical site, while in the other two series most patients had urinary tract locAL. These differences could be related to referral bias.
The second difference is the κ:λ ratio. A higher prevalence of LC κ was reported in locAL. 5 The κ:λ ratio reported in NAC and Mayo Clinic case series was 3:1 and almost 1:1, respectively. In the present study, we observed a higher prevalence of λ, with a κ:λ ratio of 1:3. This is comparable to sysAL and nicely fits with the evidence for a higher inclination of λ LC to form amyloid fibrils.
Our cohort also showed an organ type-specific variation of the κ:λ ratio for locAL, for example, almost all patients with CNS involvement presented with aLC λ, while in lymph nodes locAL aLC κ was more frequent. These observations confirm and extend previous findings from small case series of CNS and lymph node locAL. 11,22 Thus, variations of κ:λ ratios in different case series may be also related to the variable compositions of anatomical sites. Curiously, the preferential manifestation of either λ or κ in different tissues and organ sites has also been described in sysAL. 23,24 Recently, the Mayo Clinic group showed a higher frequency of LV2-14 in GI locAL, suggesting a link between IGVL usage and organ involvement, similar to sysAL. [25][26][27][28] In support of these findings, λ was also very common in our series of patients with GI locAL (83%).
For the first time, we evaluate factors that might affect local progression. In the entire cohort, local progression seems to be not affected by anatomical site, aLC isotype and presence of autoimmune system dysregulation. A concomitant MC and/or abFLCR matching the aLC has also no impact on locAL-PFS. Notably, a MC and/or abFLCR were more frequent in patients with a cAD and present in almost 50% of patients with an ANA titer >1:640. We suppose that the higher prevalence of MC than in the general population 29 is related to the high frequency of autoimmune system dysregulation observed in these patients. Indeed, a higher risk for developing secretory active B-cell lymphomas has been described. 29  to detect the clone in most cases, while more sophisticated approaches for clonality assessment, for example, in situ hybridization and molecular pathological analyses, are more effective. 5,11,13 Since these latter techniques were not systematically used in our study, it is possible that we identified only cases with more extended B-cell clones, while cases with more subtle infiltrations could have been missed for example, due to sampling errors (note the correlation with sample size) or technical issue (eg pure DNA quality, below detection limit of IGH-PCR). Indeed, it is reasonable that an amyloidogenic clone is always present at amyloid deposition site, even if it cannot be identified. Larger biopsies may be useful for a better characterization of local cellular infiltrate. However, they are not necessary for the current management of locAL.
Several studies described a clonal lymphoplasmacytic infiltrate within the spectrum of MZL in small case series with locAL. 9,10,30,31 Interestingly, in our study five patients with MZL or MALT presented with a B-cell lymphoma infiltrate at the site of amyloid deposition. Thus, a concomitant lymphoma can rarely cause locAL. However, a larger study revealed that the infiltrate is currently best classified as "localized B-cell neoplasia of undermined significance" in the majority of cases. 13 Thus, like in sysAL, the amyloidogenic clone is usually small and has not the full characteristics of a well-defined malignant disease. 4 The B-cell clones in locAL might have other not yet known biologic characteristics that affect the outcome, as already observed in sysAL. [32][33][34][35][36] The etiology of extra-nodal B-cell clones is currently unknown.
Chronic antigen exposure and autoimmune stimulation are putative explanations. 37 This hypothesis is intriguing and could explain the high prevalence of cAD in locAL, as the presence of hypergammaglobulinemia in some cases. Notably, the local clone and cAD were more frequently detected in skin locAL. Another possible factor of chronic B-cell stimulation could be cigarette smoking in lung locAL, as the prevalence of smokers was higher in these patients. However, this finding could be hampered by a pre-diagnostic bias, as smokers are more likely to get lung imaging with the consequent discovery of suspected focal lesions. The hypothesis of chronic B-cell stimulation seems to be also supported by the IGLV usage in locAL, which is similar to normal B-cell population and by the finding of inflammatory cells, especially MGC, around the amyloid deposits. Curiously, our data from samples of patients with lung locAL showed a possible protective role of the inflammatory infiltrate and MGC against local progression. These cells might interact with amyloid deposits. 5 Our study has some limitations. Due to its retrospective nature, a complete rheumatologic assessment was not available in all patients. Data about the cellular infiltrate at amyloid deposition site were present in half of cases. The clonality assessment of the infiltrate was performed with different techniques. Finally, the decision of treatment strategy (ie, watch and wait approach vs treatment) was often taken by local physicians before our evaluation.
In conclusion, our study shows that the local B-cell clone is one main driver of local progression. Response to treatment resulted in longer locAL-PFS. It is noteworthy that effective treatment approaches, as surgery and radiotherapy, result in removal and direct elimination of the B-cell clone. The removal of amyloid deposits would be another interesting therapeutic strategy. However, the effectiveness of the currently tested ant-amyloid therapies has not be confirmed so far. Finally, MGC and a local inflammatory response may be protective. However, our knowledge on the pathophysiology and biology of the B-cell clone in locAL is still poor. It is likely that, as in sysAL, the local B-cell clone harbors peculiar characteristics which might affect clinical manifestation and outcome. We hope that our work will push further basic and clinical research on locAL, in order to first improve our understanding of this mysterious disease and consecutively the management of our patients.