CytoJournal BioMed Central

Background Endosonography (EUS) guided FNA is a relatively new imaging modality which is increasingly used for sampling deep-seated lymph nodes in the diagnosis and staging of various malignancies, both primary as well as metastatic. It is also useful for staging of non-Hodgkin's lymphoma as well as diagnosing recurrence. The diagnosis of leukemia on FNA samples from deep-seated lymphadenopathy poses an even greater challenge. Hairy cell leukemia (HCL) is an uncommon, but distinct, lympho-proliferative disorder of B cell origin. It usually affects the spleen and bone marrow and uncommonly involves lymph nodes. There are only a few cases reported where HCL was diagnosed on FNA specimens. Case presentation We report the first case of HCL accurately rendered on EUS-FNA samples. Conclusion This report underscores the concept that the presence of a cytopathologist in the endoscopy suite plays an important role in providing accurate diagnoses of lymphoid lesions biopsied with EUS-FNA.


Background
Fine needle aspiration biopsy (FNAB) has established its role in the preliminary diagnosis and planning of therapy for lesions in organs like thyroid, breast, lymph nodes and even the skull [1]. It, however, plays a limited role in detection of bone lesions. This may be attributed to lack of experience of the cytological appearance of bone lesions, which in turn, is due to difficulty associated with aspiration [2].
For long, widespread usage of FNAB as a diagnostic tool has been impeded because of the requirement of considerable training and experience of the cytopathologist as well as limited clinical information, false negative and false positive diagnoses and the overlap of cytological features in benign and malignant lesions. Often diagnosis of benign lesions cannot be made with certainty. Precise classification of a tumor is difficult on the basis of FNAB alone and histopathological confirmation is frequently required. However, on the other hand, FNAB of bone lesions has its own advantages of being simple, safe, and inexpensive and a quick outpatient procedure. It can also be repeated at different sites in case of inadequate material being aspirated [3].
This study was undertaken to assess the accuracy of FNAB and histopathological correlation in the diagnosis of bone lesions. Special emphasis was given to understanding its limitations and diagnostic aberrations. Analysis of the discordant cases was done to determine the source of diagnostic errors. A review of literature was done to compare the results with those of previous workers.

Materials and methods
This study was conducted by the department of Pathology in conjunction with the department of Orthopedics, Moti Lal Nehru Medical College, Allahabad, India at the tertiary referral center: the Swaroop Rani Nehru Hospital. 91 cases were referred for FNAB in clinically suspected bone lesions.
A detailed clinical workup, including radiological assessment (X-ray and/or CT scan), was done prior to FNAB. The site of aspiration was approached through the shortest distance with radiological guidance. Aspirates were obtained using disposable needles (22-24 G) attached to a disposable 10 ml plastic syringe. A Cameco-type syringe holder was used where necessary. Under aseptic conditions, the needle was introduced into the lesion and after maintaining a negative suction pressure, multiple quick oscillations in different directions were made till some material was seen in the hub of the needle. Then the needle was withdrawn after releasing the negative pressure gently. One to two more passes were made into the lesion from different sites to ensure adequate sampling.
The physical nature of the aspirate was noted as fluid, pus, blood, caseous material and tissue bits etc. to be processed accordingly. Contents of the needle were blown on clean glass slides and the smears were made immediately. A few smears were quickly fixed in 95% ethanol for Hematoxylin and Eosin staining while the remaining smears were air dried and fixed in methanol for May Grnwald Giemsa staining. If sufficient material was left after preparing the smears, cell blocks were also prepared. In addition to the above mentioned routine stains; cytochemical stains like reticulin (Gomori's), alkaline phosphatase, Periodic acid Schiff (PAS) with without diastase, mucicarmine and immunohistochemical (IHC) stains were employed to support the diagnosis wherever necessary.
Cytodiagnostic light microscopy was embarked upon; all the smears were meticulously interpreted by two experienced cytopathologists. Accordingly, the smears were categorized as "benign", "malignant", "suspicious" and "inadequate\inconclusive". There was no perfect or absolute morphological feature of cancer, which when present unequivocally, meant that the cell is cancerous or when absent means that there was no cancer: however certain features, when taken in their totality and keeping in view the clinico-radiological findings enabled the cytopathologist to divide the cytologic findings into benign, suspicious and malignant. The smears were categorized as "suspicious" when the specimen was hypocellular and a few neoplastic cells were present or the cytological features of malignancy could not be ruled out conclusively and "inadequate\inconclusive" category was assigned when the specimens were extremely paucicellular or blood mixed to an extent that all other elements were obscured. The criteria used for labeling the aspirates as benign and malignant are depicted in Table 1. Wherever possible, an attempt was made to render an exact cytological diagnosis.
For histopathological examination, tissues were embedded in paraffin blocks, sliced into 2-3 micron sections and stained with routine Hematoxylin and Eosin staining and examined in a double blind fashion by two pathologists. In case of discrepancy, the opinion of a third pathologist was taken and two concordant diagnoses were treated as final.
The findings of FNAB were correlated with the histopathological diagnosis. Taking histopathology as the "gold standard" the diagnostic indices were calculated in terms of true and false positive, true and false negative, sensitivity, specificity, predictive values and accuracy test to support our study design. Calculation of these values was based on cases interpreted as diagnostic on histology, excluding both the inadequate\inconclusive smears as well as suspicious category.
A step-wise approach to FNAB diagnosis of bony lesions is given as follows: Step 1: Establish category of clinical presentation A patient may present with a bony mass/es under the following clinical scenarios: (i) Routine medical check-up, (ii) Bony pain, swelling or discharging sinus (iii) Known malignant cases. Important relevant data include age of the patient and site of involvement. Radiologic correlation is mandatory.
Step 2: Establish category of radiologic findings In many instances, a preoperative diagnosis can be achieved with a high degree of accuracy based on noninvasive imaging techniques and close clinical correlation. FNAB is useful in defining those lesions without characteristic imaging appearance. Lists of entities along with the cytologic and radiological findings are given as a working guide. [see Additional file 1] Step 3: Establish nature of cytohistologic findings.
Usual cytological findings are summarized in the "additional table".
Step 4: Further confirm nature of cytohistologic findings The initial cytologic assessment is crucial as it forms the basis upon which ancillary tests are ordered; the results of which should be interpreted in the larger context of the case. Special stains and IHC may be helpful. A whole battery of antibodies is available for the comparative immunohistochemical study of primary and metastatic bone tumors, specially utilizing cell block preparations. The two major diagnostic issues are (i) whether the cells are malignant or benign? (ii) what is the histogenesis of the malignant cells? For example osteosarcoma exhibit strong positivity for Vimentin, variable for Actin/Desmin and S-100, if chondroid differentiation is present.
Step 5: Establish final diagnosis based on multidisciplinary approach Close clinicopathological correlation is mandatory for enhancing the yield of FNAB diagnoses and the reduction of indeterminate reports.
A diagnostic algorithm for FNAB diagnosis of long bone lesions is given in Fig. 1, spine in Fig. 2 and skull in Fig. 3.

Results
FNAB categories of 91 cases are summarized in Table 2 A diagnostic algorithm for FNAB diagnosis of long bone lesions spine in Fig. 2 and skull in Fig. 3
In 48 benign lesions, discordant diagnoses were made in 3 cases. They included 2 cases of chronic granulomatous bone disease which, on FNAB, appeared to be metastatic adenocarcinoma, the discrepancy was due to macrophages on cytology being misinterpreted for mucin rich cells. In retrospect, PAS staining was found to be negative in these cases and a diagnosis of metastatic adenocarcinoma was refuted. In addition, a case of giant cell reparative granuloma was wrongly diagnosed as giant cell tumor on cytology (Fig. 10).
Of the 17 malignant lesions, 14 were diagnosed as malignant, 2 as "suspicious" and 1 as benign on cytology. All the inadequate/inconclusive smears were found to be benign on histology. The single discordant diagnosis was seen in a case where a diagnosis of osteochondroma on cytology was found to be an osteosarcoma on histopathology. (Fig. 11)

Discussion
Martin and Ellis first applied this technique to the diagnosis of bone lesions in 1930 [4]. Agarwal et al [5] and Layfield et al [6] have done pioneering work in describing the diagnostic accuracy and clinical utility of fine needle aspiration cytology in the diagnosis of clinically suspected primary bone tumors. As presurgical chemotherapy has become the standard treatment for osteosarcoma, FNAB has gained importance in recent years as an appealing diagnostic method [7].
Several published series have yielded overall accuracy values ranging from 51% to 100% (  [5] reported the FNAB findings in 226 cases of bone tumors. A specific morphologic diagnosis on FNAB was possible in 159 cases with one false positive and 29 false negative reports. Giant cell tumor (32%) and Ewing's sarcoma (22%) were the most common bone tumors encountered. In their series, the overall sensitivity and specificity was 86% and 94.7% respectively. The positive predictive value was as high as 99.4% while the negative predictive value was 38.3%. They reported that the diagnosis of malignant tumors was more accurate with positive predictive value of 99.2% [5]. In our series, the only false negative case was that of osteosarcoma which was misinterpreted as osteochondroma. On review, it was found that paucicellular material on aspiration, erroneous cytological interpretation of cartilaginous components and bony trabeculae, along with lack of clinico-radiological correlation was the reason for error and such a smear should have been ideally kept under "inadequate\insufficient" category.
We encountered 2 false positive cases. One case of chronic granulomatous bone lesion was erroneously diagnosed as metastatic adenocarcinoma with unknown primary. On review, this case showed a few large bizarre cells filled with mucin, which was subsequently found to be macrophages. Detailed clinical history, examination and radiological findings were not available at the time of FNAB. It was found that, when in doubt, adjunct stains like cytochemical and immunocytochemical markers helped in A diagnostic algorithm for FNAB diagnosis of long bone lesions skull  A diagnostic algorithm for FNAB diagnosis of long bone lesions spine in Fig. 2 reaching a diagnosis. The other case was that of a giant cell tumor, from an elderly male with a lytic lesion of the distal humerus, which was misdiagnosed as a sarcoma (not otherwise specified). Radiological information was again not available at the time of the diagnosis. The cause of false positive diagnosis was an interpretive error where benign cells were misinterpreted as malignant. This further underlined the importance of clinico-radiological correlation in cytology. In our series, only one metastatic malignancy was found and it was correctly diagnosed. An appropriate diagnosis of a metastatic lesion by FNAB has been reported to facilitate either non-operative management as well as contemporary surgical reconstructive techniques [30].
We also analyzed the diagnostic limitations of the technique and specimen adequacy in our study group: "inad-equate\inconclusive" smears were 10.9%. Most of these cases were osteosclerotic and fibro-osseous lesions due to frequent dry taps and inconclusive smears. FNAB has a limited role in diagnosing these lesions. However, an experienced aspirator (preferably the cytopathologist, as in our series), correct aspiration technique and proper radiological evaluation to locate the most appropriate site for adequate sampling may minimize chances of inadequate material being aspirated. This failure rate was consistent with rates of 1.4 -33% reported by previous investigators [30][31][32].  2  ----2  Giant Cell Tumor  17  16  16  -1  -Osteomyelitis  3  3  3  ---Osteochondroma  4  2  2  2  --Aneurysmal Bone Cyst  3  3  3  ---Osteoid Osteoma  2  ----2  Chondroma  2  2 Plasmacytoma  2  2  --2  -TOTAL  65  49  35  3 17 10 The separation of low-grade chondrosarcoma from enchondroma (chondroma) is an important issue but since there were only 2 cases in this series, this issue could not be addressed. Layfield et al have divided chondrosarcoma into three grades depending on the proportion of chondroid and myxoid substance as well as the degree of anaplasia [6]. Similarly, Rinas et al, in a recent report, have elucidated the difficulties in sampling errors and diagnosis of dedifferentiated chondrosarcoma [33].
In recent years, cytogenetics has been helping investigators to understand the genesis of the various bone lesions. This field is still in its infancy but cytopathologists, the world over, now recognize the fact that presence of certain chromosomal aberrations worsens the overall prognosis and survival post-therapeutic intervention in few bone tumors. For example, gain of chromosome 8q23 and CDK4 alone or together with MDM2 is associated with poor prognosis in osteosarcoma [34]. Similarly rearrangement in band 8q21 is detected exclusively in aggressive chondroblastoma [35]. More and more such associations are being discovered daily in research labs world over. However, cytogenetics was not applied in the cases under the present study but their importance, as an additional  investigative tool in future cannot be understated and a future study is planned to explore its relevance.

Conclusion
FNAB is a simple and economical technique that can be performed as an outpatient procedure, reducing patient hospitalization and lowering the overall cost of patient care. Complications are few and multiple specimens can be obtained without increased morbidity. Treatment with radiation and/or chemotherapy can be initiated without any delay. In current orthopedic oncology practice, surgeons need to know the type of malignancy present. Operative approaches as well as the use of preoperative chemotherapy and radiation therapy depend on the type of malignancy diagnosed.
When sampling is adequate and radiological findings are available, FNAB of bone is a highly accurate diagnostic technique. Inflammatory conditions, benign non-fibrotic bone lesions as well as primary and metastatic malignant tumors can be correctly diagnosed. If bony lesions appears to be fibrotic and difficult to needle with the FNAB technique, a core or open biopsy may be performed. A definitive pathologic interpretation should never be rendered if diagnostic material is inadequate or radiologic information is not compatible. Therefore, radiologists, cytopathologists, and orthopedic surgeons must work together for optimal results to avoid unsatisfactory smears. We conclude that considering the overall advantages and cost-analysis, FNAB may be suggested as the initial method of choice for evaluation of bony lesions in most clinical settings, especially in resource challenged countries. The final choice should, however, be decided on the basis of the working clinical diagnosis and the institutional/personal experience. RM participated in the analysis of data and helped to write the manuscript VO participated in the acquisition, analysis and interpretation of data.

Epithelioid granuloma of tubercular osteomyelitis
PS participated in the acquisition, analysis and interpretation of data.