Diffuse lesion and necrosis tied to poorer prognosis of interdigitating dendritic cell sarcoma: cases report and a pooled analysis

Interdigitating dendritic cell sarcoma is a neoplastic proliferation of interdigitating dendritic cells and no therapeutic consensus exists. This study aimed to investigate the prognostic impacts of tumor lesion, cellular atypia, mitosis and necrosis on the interdigitating dendritic cell sarcoma. Case reports and pooled analyses were designed to explore the relationships. One case was a 40-years old man with localized lesion, moderate to notable cellular atypia, 30 mitoses per 10 high-power fields and no necrosis and the progression-free survival was longer than 20 months. The other case was a 62-years old woman with diffuse lesion, notable cellular atypia, less than one mitosis per 10 high-power fields and diffuse necrosis and the progression-free survival was shorter than 1 month. Cellular atypia and mitosis had not any relationship with survival. Compared with localized lesion, diffuse lesion presented a 2.92-fold risk of progression (HR = 2.92, 95% CI 1.01, 8.51) and an 8.79-fold risk of death (HR = 8.79, 95% CI 1.86, 41.64). Diffuse necrosis presented a 4.39-fold higher progression risk (HR = 5.39, 95% CI 1.78, 16.29) and a 5.37-fold higher death risk (HR = 6.37, 95% CI 1.46, 27.86) than focal or no necrosis. Diffuse lesion and diffuse necrosis were indicators of poorer prognosis and the clinical application should be warranted in further studies.

smashed white-grayish lymph nodes and the section showed solid, white-grayish, focal black, rough and diffused necrosis (Fig. 1B).
ISH and PCR. Detection of Epstein-Barrvirus-encoded small RNA was negative for two cases by in situ hybridization. T and B cell polymerase chain reaction (PCR) clonality studies were not rearrangement. Braf with V600E mutation was not detected.
Electron microscopic analysis. The tumor cells had mutual interlaced long finger bumps and the borderline was unclear in the two cases. There were small amount of rough endoplasmic reticulum in cytoplasm. Nucleus had dents. There was no a typical bridge (Fig. 4A,B).
Patients with diffused lesion had median PFS of 3 months, in contrast, patients with localized lesion had not reached median PFS (Fig. 6A). The difference was significant. Cellular atypia and mitosis were not associated with PFS (Fig. 6B,C). Median PFS of diffuse necrosis was 3 months but patients with no necrosis or focal necrosis had not observed median PFS, the difference being significant (Fig. 6D). In multivariate analysis, diffuse lesion presented a 2.92-fold risk of progression (HR = 2.92, 95% CI 1.01, 8.51) and an 8.79-fold risk of death (HR = 8.79, 95% CI 1.86, 41.64) ( Table 2).
Diffuse tumor lesion was associated with a shorter OS significantly (Fig. 7A). Patients with localized lesion did not reach median OS, but patients with diffuse lesion had a 9-month OS. Cellular atypia and mitosis were not related with OS (Fig. 7B,C). Patients with no or focal necrosis did not get median OS, compared with patients with diffused necrosis having a 9-month median OS (p < 0.001) (Fig. 7D). Diffuse necrosis was associated with a 5.39-fold high progression risk (HR = 5.39, 95% CI 1.78, 16.29) and a 6.37-fold high death risk (HR = 6.37, 95% CI 1.46, 27.86) ( Table 2).

Materials and Methods
This study was approved by Beijing Shijitan Hospital Institutional Review Board (IRB). Written informed consent was obtained from all patients. No identifying information or image was published. All of the methods were in accordance with the particular regulations.
Study design. This study design consisted of a case report and a pooled analysis.
Pooled analysis. This study chose the databases of PubMed and CHINAINFO. PubMed was accessed through NCBI and CHINAINFO was accessed through WANFANG DATA.
Search term was "interdigitating dendritic cell sarcoma". The publication date ranged from Jan. 1, 1980 to Mar. 1, 2016. The language covered English and Chinese. Included article type was original research and the excluded types were reviews, meeting abstract, commentaries or re-analysis of data. All of the include IDCS cases should had pathlogical diagnosis in the publication. The endpoint outcome was progression-free survival (PFS), and overall survival (OS). The extracted data were age, gender, tumor lesion, cellular atypia, mitosis, necrosis, PFS and OS.
126 articles were obtained in PubMed and 53 articles were obtained in CHINAINFO. We excluded 42 articles in PubMed and 11 articles in CHINAINFO without IDCS cases. We reviewed all of the rest articles and found 34 cases in PubMed and 17 cases in CHINAINFO with histopathology data and outcome data.
Our two IDCS cases came from Beijing Shijitan hospital. The confirmed diagnosis covered by histological review of morphology in slides stained with hematoxylin and eosin, detection of a wide immunohistochemistry panel and application of electron microscope. The two cases were included in the pooled analysis.
Case report. Case 1: a 40-years old man visited the outpatient clinic presenting a two-week history of a painless mass in his left inguinal. The patient had foot nail removal for paronychia caused by trauma five years ago. No skin moles and tumor was found at lower limbs. The patient had no family cancer history and was in good health before. Ultrasound showed one low echo node in the right neck, the size being 0.9 × 0.2 cm; some low echo nodes in the left neck, the biggest size being 0.5 × 0.3 cm; some low echo nodes in the right inguinal, the biggest size being 0.8 × 0.3 cm; some low echo nodes in the left inguinal, the biggest size being 1.3 × 1.0 cm. All of the nodes had clearly border and regular contour. Pelvic CT (Fig. 8A) showed the lymph nodes aside the left external iliac blood vessels enlarged and the biggest one was 1.8 × 1.4 cm, but there was no enlarged lymph node in the right external iliac blood vessels and retroperitoneal. There was no abnormality in the liver and spleen. PETCT (Fig. 8B) displayed a high bone salt metabolism in right coracoid and left sacroiliac joint, indicating a benign lesion. Blood and bone marrow aspiration were normal.
Case 2: a 62-years old woman presented cough and hard breath for 3 months. Superficial lymph nodes were not touched and blood test was normal. Chest CT (Fig. 8C) scan showed: multiple lymph nodes enlarged in right hilar and mediastinal, multiple nodules with calcification in the upper lobe of right lung and some patchy enhancement spot in the middle of the upper lobe, left lower lobe and secondary adjacent bronchiectasis. Abdomen CT (Fig. 8D) scan showed a clear borderline mass in the right kidney, with the size of 4.4 × 4.1 × 5.8 cm, and a slightly larger spleen. Mediastinal lymph node was biopsied and pathological diagnosis was tending to be metastasis of poorly differentiated carcinoma with a large number of necrosis. The patient had two cycles chemotherapy of paclitaxel and nedaplatin but the disease progressed. The patient had mediastinal lymph node biopsy again.
Definition. Diffuse lesions means the lesion involved other organs outside the primary site. Focal necrosis is small, usually no more than 0.3 mm, and the margin is relatively regular, cannot be seen in the gross appearance; diffuse necrosis is large, usually more than 0.5 mm, and has irregular margin, often be seen in the gross appearance, it is also been called "map-like" necrosis.
Immunochemistry. All immunohistochemical stains were performed on the Ventana Benchmark automated staining system using 4 mm paraffin tissue sections. The primary antibodies used in this study including CD68, S100,CKpan, cam5.     follicular DCs and IDC. IDC reside in t-zones in the peripheral lymphoid tissues, such as the paracortex and deep cortex of lymph nodes. Dendritic neoplasms are rare hematological malignancies with less than 1% occurrence in the lymph nodes or soft tissues. IDCS can occur in lymph node, nasopharynx, small intestine, mesentery, spleen, testis, skin, tonsil, bladder, eyelid, uterine cervix and so on 6,9,[13][14][15] . Usually, IDCS had an indistinct border, astoriform pattern of spindled to ovoid cell and abundant slightly eosinephillic cytoplasm. Small to large multinucleate cells appeared with distinct nucleoli. IDCS had a moderate cytologic atypia and a low mitotic rate. Necrosis was less seen 4 . In our overview, we found that IDCS had epithelioid cells and mononucleated or multinucleated tumor giant cells, a high mitotic rate, diffuse necrosis and granuloma 5,16,17 . Patients usually present with painless lymph node enlargement or extranodal masses. Systemically atypical symptoms include fever, weight loss, fatigue and night sweats. IDCS were often misdiagnosed as carcinoma, melanoma, sarcoma, and so on [18][19][20] . There was diversity in histopathological morphology. Immunohistochemically, there is no specific marker for IDCS. Therefore, the diagnosis of IDCS was difficult and achieved by excluding other diseases. IDCS cells were consistently positive for S100 protein and vimentin, weakly positive for fascin, CD68, lysozyme and CD45. The tumor cells were negative for CD21, CD23, CD35 and CD1a, CKpan, cam5.2, EMA, HMB45, Melan A, CD3, CD20, Desmin, SMA, and so on. However, in case 2, the tumor cells expressed CKpan focally, which needed further investigation. Ultrastructurally, the cells demonstrated complex features of interdigitating cells. Birbeck granules and melanosomes were not seen from other studies 4, 18 .    Patients have variations in prognosis (some patients received operation and chemotherapy and radiotherapy, but PFS was no more than 1 month, some patient had no treatment after surgery but PFS lasted for years) 15 . The extreme rarity of IDCS impeded researches about the prognostic factors for the diverse prognosis. Pas 21 et al. reported necrosis and mitosis were unconfirmed prognosis factors. Therefore, we proposed a pooled analysis for IDCS for the histopathological prognosticators. Our analysis showed median PFS of diffuse necrosis was 3 months but patients with no necrosis or focal necrosis had not observed median PFS, with a significant difference. Caner Saygin 22 et al. studied 25 cases of IDCS and reported that necrosis was not an indicator of poor prognosis. It was not conflict with this study, because we divided the necrosis into focal necrosis and diffuse necrosis and the diffuse necrosis indicated the poorer prognosis.

Discussion
Our analysis indicated high mitotic count (≥5/10 HPF) and the nuclear atypia were not associated with the poor prognosis the results were in accord with previous report 21 . We also found median age has not any significant differences between tumor lesion, cellular atypia and necrosis. However, the patients with nuclear mitosis <5/10HPF were younger than the patients with nuclear mitosis ≥5/10HPF. The significance is needed to be confirmed further. Patients with diffused lesion had a shorter PFS. These results indicated diffuse lesion had a worse prognosis and were identical with the reported in literature 5,23 .
However, there are some limitations of our analysis. Our data stemmed from the published literature and we could not obtain complete clinical, pathological and follow-up data for some patients. It is hard to compare IDCS patients with uniform treatments. Therefore, it was necessary to implement a consolidated and well-designed prospective design in further study.
In conclusion, we found that the diffuse lesion and diffuse necrosis of IDCS was associated with a worse biological behavior and prognosis, and the cellular atypia and mitosis had no distinct relevance with the biological behavior and prognosis of IDCS.