Pathologically confirmed oesophageal involvement in idiopathic multicentric Castleman disease mimicking early oesophageal cancer

DearEditor, Castleman disease (CD) is a very rare lymphoproliferative disease that presents in two distinct clinical forms: the localized form (unicentric CD) and the multicentric form [multicentric CD (MCD)] [1]. Only about 1500 cases of MCD have been identified in Japan by the Japan Intractable Diseases Information Center. Here, we describe an idiopathic MCD patient who presented with an oesophageal lesion resembling those of early-stage oesophageal cancer. The patient was a 65-year-old man who had a 10-year history of bilateral lung infiltrates, multiple lymphadenopathies, anaemia, and elevated serum CRP and IgG levels of unknown aetiology. He was referred to our hospital, and idiopathic MCD was diagnosed on the basis of these clinical manifestations as well as histopathologic findings of marked plasmacytic infiltration in the interfollicular region of the right inguinal lymph node, compatible with plasma cell-type CD (Fig. 1A–D); negative findings for HHV8 infection; and elevated serum IL-6 levels (26.3 pg/ml), according to the international diagnostic criteria [2]. Endoscopy for screening of anaemia revealed an 0-IIc-like lesion (∼40 × 10 mm) in the middle part of the oesophagus. The white light observation showed a slightly thick impression (Fig. 1E). The narrow-band imaging observation revealed mainly Type B1 in the magnifying endoscopic classification of the Japanese Esophageal Society [3] (Fig. 1F and G). Iodine staining was not distinguishable. Since the biopsy did not allow us to distinguish between intraepithelial neoplasia and inflammatory atypia, we performed endoscopic submucosal dissection for this lesion. The specimen size was 58.5 × 20.3 × 5.4 mm (Fig. 1H), and the histopathologic findings showed follicular hyperplasia with marked plasmacytic infiltration in the interfollicular region (Fig. 1I and J). Immunohistochemical findings showed CD20-positive B cells within the follicles (Fig. 1K) and CD138-positive plasma cells outside the follicles (Fig. 1L). These findings were compatible with plasma cell-type CD as well as those found in a previously resected right inguinal lymph node (Fig. 1A–D). Thus, the oesophageal lesion, which was completely resected by endoscopic submucosal dissection, was considered to be one of the primary organ involvements of idiopathic MCD. Although the patient had anaemia, elevated CRP and IgG, bilateral lung infiltrates and lymphadenopathy, he had had no obvious clinical symptoms for the previous 10 years. Therefore, he chose careful follow-up without any specific treatment, such as CS, immunosuppressants or tocilizumab, for the idiopathic MCD. The patient presented in this report gave his written informed consent prior to his inclusion.

Fig. 1

Histopathologic features of lymph node, endoscopic and histopathologic features of esophageal involvement in idiopathic multicentric Castleman disease (A and B) Histologic findings of the right inguinal lymph node revealed marked plasmacytic infiltration in the interfollicular region (HE staining, the bar indicates 200 micro m in A, and 50 micro m in B). (C and D) Immunohistochemistry of the right inguinal lymph node showed CD20-positive B cells within the follicles (C, the bar indicates 200 micro m), and CD138-positive plasma cells were observed outside the follicles (D, the bar indicates 200 micro m). (E) On endoscopy, an 0-IIc-like lesion (approximately 40 x 10 mm) with slightly thick mucosa was observed in the middle part of the esophagus (indicated by arrowheads). (F) Narrow-band imaging (NBI) showed a brownish area in the center (indicated by arrowheads). (G) Magnifying NBI observation revealed mainly type B1 according to the endoscopic classification of the Japanese Esophageal Society. (H) The size of the resected esophageal specimen was 58.5 x 20.3 x 5.4 mm. (I and J) Histologic investigation of the resected esophageal specimen revealed follicular hyperplasia with marked plasmacytic infiltration in the interfollicular region (HE staining, the bar indicates 200 micro m in I, and 50 micro m in J). (K and L) Immunohistochemistry of esophageal specimen showed CD20-positive B cells within the follicles (K, the bar indicates 200 micro m), and CD138-positive plasma cells were observed outside the follicles (L, the bar indicates 200 micro m).

Open in new tabDownload slide

To the best of our knowledge, this is the first report describing pathologically proven oesophageal mucosal involvement in MCD. A recent report revealed that the most common clinical presentations besides lymphadenopathy included fatigue (61.7%), pulmonary involvement (42.7%) and splenomegaly (41.8%), whereas gastrointestinal manifestations such as decreased appetite (15.2%), constipation (5.8%) and gastric ulcer (2.9%) were relatively rare in 342 patients with MCD treated with tocilizumab in Japan [4]. A report of four cases in which the pathologic findings of gastrointestinal manifestations in patients with CD were evaluated showed that two of the patients had gastric erosions whose histopathologic findings were nonspecific inflammation, one patient presented with an abdominal mass of plasma cell-type CD that pressed on the bowel wall causing colon ulcers and one patient had concomitant rectal carcinoma [5]. These four cases suggested that the digestive tract could be secondarily involved in CD patients. On the other hand, only one patient with pathologically confirmed hyaline vascular-type CD involving the oesophageal submucosal layer was previously reported [6]; no reports have been published of CD affecting the oesophageal mucosa, as was the case in our patient. In conclusion, our case indicated that idiopathic MCD could involve the oesophageal mucosa, mimicking early oesophageal cancer. Screening of the gastrointestinal tract including the oesophagus might be necessary for patients with idiopathic MCD.

Acknowledgements

We thank Dr Flaminia Miyamasu, Medical English Communications Center, University of Tsukuba, for language review of this letter.

Funding: No specific funding was received from any funding bodies in the public, commercial or not-for-profit sectors to carry out the work described in this manuscript.

Disclosure statement: The authors have declared no conflicts of interest.

References