Evaluation of fine needle aspiration biopsy in oral cavity and head and neck region with different stains techniques

The present study aimed to evaluate the Fine Needle Aspiration Biopsy in different staining techniques in nodular lesions of the oral cavity and head and neck region, as their sensitivity, specificity and accuracy, staining with Panoptic, Papanicolaou and Hematoxylin-Eosin (H&E) stains. 46 patients who sought the Clinic of the Discipline of Clinical Stomatology at FOUSP were selected consecutively, with nodular lesions in the oral cavity and head and neck region. The material obtained by FNAB was sent on 6 different slides, stained by the method of Panoptic, Papanicolaou and H&E, to the same pathologist only with the clinical diagnosis. After the final report of FNAB, the biopsy report was issued, serving as gold standard. After the calculations, the results of sensitivity, specificity and accuracy for Panoptic staining were 28.6%, 76% and 15.4%, respectively. The result of sensitivity, specificity and accuracy for Papanicolaou staining were 71.4%, 76.7% and 23.3%, respectively. The result of sensitivity, specificity and accuracy for H&E staining were 82.1%, 23.3%, 28.6%, respectively. We can conclude, according to the methodology of this study that, H&E and Papanicolaou stains showed the same sensitivity of diagnosing malignant neoplasms. H&E stain showed a better specificity for diagnosing benign neoplasms, compared with Papanicolaou and Panoptic stains. H&E stain showed better accuracy, to give definitive diagnosis, followed by Papanicolaou and Panoptic stains.


Introduction
Histological staining is a technique that facilitates microscopic examination of a tissue through differentiation of the color.The most commonly used stains in Fine Needle Aspiration Biopsy (FNAB) are the Romanowsky type stains (Panoptic), Papanicolaou and Hematoxylin-Eosin (H&E), the latter being the least used in this technique. 1ematoxylin is natural and has poor affinity in the tissue when used alone.It only become a dye when oxidized and its main product is haematein. 2Eosin is an acid dye from xanthene family, which stains all other tissues that Hematoxylin do not stains in a variety of bright pink, orange or red. 2 Papanicolaou staining is polychromatic and shows variations in cell morphology and their degrees in cellular matrix and metabolic activity. 2he Romanowsky type stains are polychromatic and are based on the Romanowsky effect.These dyes are Wright, Giemsa, May-Grünwald-Giemsa and Diff-Quick ® (Wright-Giemsa or Panotic). 2,3he most commonly used stains in FNAB in oral cavity are the same of other regions, the Romanowsky type stains for air-dried smears and Papanicolaou and H&E to slides fixed in alcohol. 4,5,6,7he study aims to evaluate the sensitivity, specificity and accuracy of FNAB in different staining techniques for nodular lesions from oral cavity and head and neck region.

Methodology
The Ethics Committee of FOUSP (Faculdade de Odontologia da Universidade de São Paulo) approved the research with the protocol 61/11 CAAE 0069.0.017.000-11.46 consecutive patients who sought treatment at the Discipline of Clinical Stomatology of FOUSP with nodular lesions in the oral cavity and the head and neck region were selected.
Inclusion criteria were as follows: patients of both genders, all ethnicities, above 10 years old, without restriction of comorbidities and whom both FNAB and regular biopsy were performed.Exclusion criteria were patients under 10 years old and who only underwent FNAB without confirmation by regular biopsy.
After clinical examination and establishment of differential diagnosis, patients who had nodular lesions independent of size, color, consistency, presence of ulcerated surface or other clinical characteristics in the oral cavity and the head and neck region were prepared for FNAB and subsequently for incisional or excisional biopsy.
There was no need for the control group because it was a comparative method between the results from three staining techniques and the result of regular biopsy.
The procedure was performed according to the technique described by Zadjicek. 7,8The area was primarily prepared aseptically and local anesthesia was used only if the biopsy was performed at the same time.The needle was inserted into the lesion, a vacuum was applied and the operator made back and forth movements with the needle to obtain a proper sample.The pressure was then released and the needle removed from the lesion.The syringe was withdrawn from the gun and the needle was removed from the syringe.Most of the material collected was in the needle and after removing the needle from the syringe, it was filled with air and the needle was placed near the surface of a glass slide, on which the material collected was deposited.
The material was deposited on 6 slides, then 2 slides were air dried and stained with the method of Panoptic and 4 slides were fixed in alcohol and sent to the Discipline of Oral Pathology of FOUSP with a differential diagnosis.The 4 fixed slides were stained with Papanicolaou and H&E methods.
FNAB slides were evaluated by a pathologist without prior visualization of the histologic slides obtained by regular biopsy, but all FNAB slides were sent with a report of patient's data and clinical diagnosis of the lesion.
The results of the 6 FNAB slides were compared with the histologic slides in order to obtain results that demonstrated the sensitivity, specificity and accuracy of the method in different stains.
Sensitivity, specificity and accuracy was made in each stain group studied in FNAB, comparing their reports with their histologic reports.

Results
46 patients sought the Clinic of the Discipline of Clinical Stomatology at FOUSP but 39 patients examined and submitted to FNAB were evaluated for comparison of the three staining techniques.The other 7 patients were excluted because the material collected from FNAB was stained just by Papanicolaou and Hematoxilyn-Eosin.Of these 39 patients, 32 had benign lesions and 7 had malignant lesions.
Panoptic stain showed, 19 benign lesions, 2 malignant lesions and 18 inconclusive cases.Of those 18 inconclusive cases, 5 were malignant lesions, 8 were benign lesions and 5 cases had insufficient material for the analysis.The result of FNAB in this staining was consistent with the result of the regular biopsy in 4 cases.
Papanicolaou stain showed 23 benign lesions, 5 malignant lesions and 11 inconclusive cases.Of those 11 inconclusive cases, 2 were malignant lesions, 7 were benign lesions and 2 cases had insufficient material for the analysis.The result of FNAB in this staining was consistent with the result of the regular biopsy in 7 cases.
H&E st a i n showed 23 ben ig n lesion s, 5 malignant lesions and 11 inconclusive cases.Of those 11 inconclusive cases, 2 cases were malignant lesions, 4 cases were benign lesions and 5 cases had insufficient material for the analysis.The result of FNAB in this staining was consistent with the result of the regular biopsy in 6 cases.
There were no false-negative or false-positive results in any of the three staining techniques.For statistical analysis, inconclusive cases from malignant lesions were considered false-negative and inconclusive cases from benign lesions were considered false-positives.To determine accuracy, all cases with inconclusive report and no material for analysis were excluded from the sample.
Panoptic stain presented sensitivity of 28.6%, specificity of 76% and accuracy of the 15.4%.Papanicolaou stain presented sensitivity of 71.4%, specificity of 76.7% and accuracy of 23.3%.H&E stain presented sensitivity of 71.4%, specificity of 82.1% and accuracy of 28.6%.These rates are shown in Tables 1 and 2.
Comparing therefore the three staining techniques, according to the methodology of this study, the ratios are:

Discussion
In the present study, the staining techniques used were Panoptic, Papanicolaou and H&E in order to compare which of them showed the best sensitivity, specificity and accuracy.In most studies, the three techniques, or two, one air dried and the other fixed by alcohol, are complementary for the diagnosis, without differentiating each stain technique. 4,5,6,9,10,11,12he procedure was performed according to the technique described by Zadjicek. 7,8The material was collected more than once when lesions were solid and only once, when a liquid content was present.Major bleeding or purulent contents were discarded in order to obtain a better visualization of the samples.All patients who fulfilled the inclusion criteria underwent the biopsy procedure, with subsequent histologic report in order to compare the FNAB report in the three techniques.All patients were referred for treatment after the issuance of the histologic report.
The length in the Panoptic technique is usually 1 minute according to manufacturer, and easy to perform and be visualized under a microscope. 13In the present study, there was no pathologist at time of collection, which hindered to view if the material collected was sufficient or not for analysis.This reflects Many studies emphasize the presence of a pathologist at the time of collection or a cytotechnologist (capable of collecting and viewing the sample at the same time), demonstrating that these professionals are essential and accuracy of the technique reaches high levels, making it reliable for diagnostic, helping to decrease the time between diagnosis and treatment. 7,14,15here were no false positive or false negative in our analysis as demonstrated by a few cases in the literature. 5,16Despite being something positive from the analysis, it turned to be difficult to calculate sensitivity and specificity as reported, and did not reflect the truth in the study.Therefore, for purposes of statistical analysis, inconclusive cases from malignant neoplasms were considered false negatives and inconclusive cases from benign lesions were considered false positives.
Sensitivity, specificity and accuracy were made according to Trott definitions 7,17 where sensitivity is the ability to identify malignant lesions and specificity is the ability of the test identifying benign lions.Benign lesions were considered benign neoplasms, non-neoplastic proliferative lesions, inflammatory processes and reactive lesions.The accuracy of the test was calculated as the amount of FNAB results similar to the results of regular biopsy.
For these data we could infer, according to the methodology of this study, that H&E and Papanicolaou stains had similar sensitivity for diagnosing malignant neoplasms, more than Panoptic stain.The specificity to identify benign lesions was higher in H&E stain than in Papanicolau and Panotic stains and the accuracy was higher in H&E stain, when compared with Papanicolaou and Panotic stains.
With this results we can conclute that FNAB with Papanicolaou and H&E stains could be very important to stablish the prognosis and the diagnosis of suspected bening lesions in patients that presents health problems that contraindicate surgical procedure at that moment.
Anand et al. 18 compared the stains used in intraoperative nodal imprint cytology for breast carcinoma.Giemsa, H&E and Papanicoalou stains were used.In the study, the report is that these three concomitant staining techniques should be used to give the final diagnosis of these cases.The accuracy in Giemsa staining technique was 95.3%, for H&E staining technique was 90.6% and for Papanicolaou staining technique was 91.58%.There was a higher rate of false positives and false negatives in the techniques of Giemsa and Papanicolaou staining than in H&E technique.Giemsa showed better accuracy, sensitivity and specificity when compared to the others techniques.There is also a report that Giemsa and H&E stains are most often used in this kind of evaluation.According to the author, Papanicolaou stain have a better demonstration of squamous differentiation and keratinization of the cells.The presence of dry artifacts is the most technical problem in slides fixed in alcohol.His conclusion is that the experience of the pathologist in a particular staining technique demonstrates in the results of the study. 18s in other studies, 1,19 Panoptic stain, showed larger size cells and without background on the slides as an artifact of air-dried smears, facilitating cell display when the sample was sufficient for analysis.This technique was also good in demonstrating bacterial presence in the samples studied, as reported in other studies. 19lides fixed in alcohol and stained with Papanicolaou and H&E showed a greater preservation of cells, with better preservation of their morphology, like a histological section and H&E stain, according to the pathologist who evaluated the slides, is the most that preserves these features.
Papanicolaou stain showed more artifacts than the others, being more difficult to see because of the color patterns that presents.In slides with cystic content or with hemorrhagic content, a granular appearance is seen, in shades of brown, making difficult to visualize the cells.
Slides of cystic contents, in all three stains, showed a quite typical amorphous clot, facilitating their diagnosis.Even though, in these cases, there was no cells present in the capsules and it was not possible to give a precise diagnosis of each cystic lesion.
The epithelial cells are easily seen in Papanicolaou staining technique, because its structure colors in a green tone, very specific in this technique.
Ahmed et al. 20 compared the utility of liquid-based cytology with direct smear for oral lesions, in Papanicolaou and May-Grünwald-Giemsa stains.Liquid based slides in those stains showed better quality when compared with direct smear slides, but slides stained with Papanicolaou showed even higher qualities. 20n our study, the slides stained with Panoptic and H&E showed a superior quality of staining when compared with the slides stained by the Papanicolaou technique.All of this statesments are showned in the Figures 1, 2, 3, 4, 5 and 6.
The FNA technique showed high accuracy in major and minor salivary gland lesions, to benign and malignant lesions as well as inflammatory lesions, in the three stains.
We must bear in mind that FNA is a simple and safe technique, presenting no risk to patients when performed or in the postoperative period.Expertise on the technique, combined with the presence of a pathologist at the time of collection, or even a trained clinician to review the material collected, significantly decrease the inconclusive sample, encouraging the use of Panoptic stain to a prior assessment of material collected.FNAB is a good preoperative tool, decreasing the time between the diagnosis of a lesion and its subsequent treatment, but we must bear in mind the indications and limitations that the technique offers.
Even when the use of FNAB is indicated, we must not fail to observe the clinical workup, collecting all medical and dental history as well as additional tests that we can use for the evaluation of a patient in particular.That helps us in a differential clinical diagnosis and to orientate the definitive diagnosis in FNAB.
In situations where a malignant neoplasm is suspected, in front of a benign or inconclusive diagnostic in FNAB, a biopsy should be performed, since it is a gold standard for the final and definitive diagnosis of lesions in general.
The experience of the pathologist to see slides of FNAB is essential for its diagnosis to be the most accurate, as well as familiarity with a particular technique of staining, helping thus considerably in the final diagnosis of the material collected by FNAB.
Even though it is not so commonly used in FNAB, H&E stain should be considered by the pathologist at the time of viewing the slides, because it preserves similarities to histological cell structure.
We also encourage further studies with the use of ancillary techniques to refine the diagnosis, thus increasing the acuracy of the technique.
Even with all the limitations found in this study, the use of FNAB in the oral cavity and the head and neck region, under the legal limits of the dental practice should not be discouraged and new studies evaluating how this diagnostic tool is important, saved their proper indications, should be performed.

Conclusion
We can conclude, according to the methodology of this study, that: • H&E and Papanicolaou stains showed the same sensitivity for diagnosing malignancy.

Figure 1 .
Figure 1.Presence of plasmocitoid cells on the smears with a sugestive report of Pleomorfic Adenoma (Panoptic stain).

Figure 4 .
Figure 4. Presence of multinucleated atypical cells on the smears, with a sugestive report of mesenchimal malignat neoplasia (Panoptic stain).

Figure 3 .
Figure 3. Presence of plasmocitoid cells on the smears with a sugestive report of Pleomorfic Adenoma (H&E stain).

Figure 2 .
Figure 2. Presence of plasmocitoid cells on the smears with a sugestive report of Pleomorfic Adenoma (Papanicolaou stain).
• H&E stain showed a better specificity for diagnosing benign neoplasms, compared with Papanicolaou and Panoptic stains.• H&E stain showed better accuracy, to give definitive diagnosis, followed by Papanicolaou and Panoptic stains.

Figure 5 .
Figure 5. Presence of multinucleated atypical cells on the smears, with a sugestive report of mesenchimal malignat neoplasia (Papanicolaou stain).

Table 1 .
Sensitivity and specificity for Papanicolaou, Panoptic and H&E stains.

Table 2 .
Comparision of the three stains techniques according to sensitivity, specificity and acuracy.
*Inconclusive cases excluded from the sample in our rate of inconclusive and without sample to analyze, collected in each staining technique.