J Oral Maxillofac 49262-267.
Surg
1991
Fine-Needle Aspiration Biopsy of Head and Neck Lesions P. MICHAEL SCHELKUN,
DDS, MD,* AND WALTER G. GRUNDY, MDt
The case records of 213 fine-needle aspiration biopsies (FNAB) of head and neck masses performed on 209 patients over a 3-year period were reviewed to assess the diagnostic accuracy and safety of this technique in comparison with surgical histologic examination. Cytologic diagnoses based on FNAB were compared with histologic diagnoses in 110 patients who underwent surgery. Based on cytology alone, 40.3% of the lesions were reported as malignant, 45.1% as benign, and 14.6% as indeterminant. A specific cytologic diagnosis was made in 85.5% of the cases. Cytologic diagnoses concurred with surgical histologic diagnoses 90% of the time. Fine-needle aspiration biopsy was found to have a false-positive rate of 0.5% and a false-negative rate of 2.3%. The sensitivity and specificity of FNAB in determining a malignant diagnosis were 81 .l% and 99%, respectively. Positive and negative predictive values were calculated at 98.9% and 82.8%, respectively. Diagnostic rate, sensitivity, and negative predictive value increased consistently throughout the study period, indicating that the diagnostic accuracy of FNAB improved with experience.
and safety in comparison with the widely accepted standard of histologic examination of surgical specimens.
Although 60 years have passed since Martin and Ellis published their now famous treatise on fineneedle aspiration biopsy (FNAB),’ the American medical community has been slow in recognizing the value of this diagnostic technique. Over the past 30 years, FNAB has been widely applied on the European continent, particularly in Scandinavia, and has proved to be a safe, rapid, inexpensive, and accurate method for establishing a diagnosis and defining treatment.2-6 Based on the published European experience and recent experience at some American centers,‘-‘* it now appears that FNAB is gaining wider acceptance among clinicians. In a retrospective review, we analyzed our experience with FNAB of head and neck masses over the past 3 years to assess its diagnostic accuracy
Materials and Methods The case records of all patients who underwent FNAB of head and neck masses at our institution between May 1987 and March 1990 were reviewed. This included biopsies of all lesions located superior to the clavicles. Over 96% of the masses were biopsied and interpreted by one of the authors (WGG). NEEDLE
In all cases, the technique used to collect the biopsy specimens was based on one previously described by Zajicek and Frable,4,10 with some minor modifications. If the lesion could be easily palpated and stabilized, a lo-mL Luer-Lok (BectonDickinson, Rutherford, NJ) syringe in a Cameco Syringe Pistol Holder (Precision Dynamics, San Fernando, CA) with a 23-gauge needle was used. If the mass was small and/or relatively superficial, however, a 23-gauge needle was attached to one
* Resident,
Department of Surgery, Rush-Presbyterian-St Luke’s Medical Center, Chicago. t Clinical Assistant Professor, Department of Pathology, The University of Alabama in Huntsville School of Primary Medical Care, and Huntsville Hospital, Huntsville, AL. Address correspondence and reprint requests to Dr Schelkun: 3714 N Wayne St, Chicago, IL 60613. 0 1991 American geons
Association
of Oral and Maxillofacial
BIOPSY TECHNIQUE
Sur-
0278-2391/91/4903-0008$3.00/0
262
SCHELKUN
AND GRUNDY
end of an intravenous extension tube and the IO-mL Luer-Lok syringe to the other (Fig 1). An assistant was required to create negative pressure within the tubing apparatus. Following anesthetization of the skin with 1% lidocaine, two or three fine-needle aspirations were generally performed at each biopsy site. After each aspiration, a drop of material was placed on a slide and a smear was prepared. Approximately half of the slides were fixed in Reagent Alcohol (Fisher Scientific, Fairlawn, NJ), and half were air-dried. The needle and syringe were then rinsed with Muccolex (Lerner Laboratories, Pittsburgh, PA), a cytologic fixative. If the material was scant, the majority of slides were air-dried. The slides were routinely stained with Diff-Quik (American Scientific Products, McGaw Park, IL), Papanicolaou, and hematoxylin-eosin stains. Special stains were used when indicated. The solution of Muccolex and the aspirate was passed through a millipore filter and placed on slides and stained as previously described. Experience dictated a modification of this biopsy technique for thyroid gland aspirates. Because of the highly vascular nature of the thyroid gland, and the propensity to aspirate mostly blood, the use of a 23-gauge or 25gauge, 5-cm Inrad (Grand Rapids, MI) aspiration biopsy needle was adopted. We found that by using the natural capillary action of the needle, a more diagnostic aspirate could be obtained. STATISTICAL METHODS
Data obtained from the review of case records were first categorized according to the regional lo-
263
cation of the biopsy site: thyroid, neck, parotid, scalp, or intraoral. We then noted both the general cytologic diagnosis (malignant, benign, or indeterminant) and the specific cytologic diagnosis, when available, for each of the lesions. Cytologic findings were compared with the results of histological examination of surgical specimens in those patients who subsequently underwent an operative procedure to determine the specific diagnostic rate and number of false-positive and false-negative diagnoses obtained by FNAB. The diagnostic rate was calculated by dividing the number of cases in which there was agreement between the cytology and the surgical histology by the total number of cases that were treated surgically. The sensitivity and specificity of FNAB for determining malignancy, as well as positive and negative predictive values, were also calculated. Sensitivity was calculated as TP/(TP + FN), for which TP equals the number of lesions identified as malignant and FN is the number of false-negatives. Diagnostic specificity was calculated as TN/(TN + FP). for which TN equaIs the number of lesions identified as benign and FP equals the number of false-positives. Positive predictive value was calculated as TP/ (TP + FP), for which TP equals the number of lesions diagnosed as malignant and FP equals the number of false-positives. Negative predictive value was calculated as TN&TN + FN), for which TN equals the number of lesions identified as benign and FN equals the number of false-negatives. In our analysis, the false-negative values used in the calculation of sensitivity and negative predictive value were determined in three different ways. The first value (FN i), or true false-negative, included ah lesions with which cytologic diagnosis was designated as malignant or suspicious for malignancy and with benign surgical histology. The second value (FN,) included the true false-negatives plus the total number of indeterminants. The third value (FN3) was defined as the true false-negatives plus the indeterminants, minus the total number of patients within the indeterminant group who were noncompliant with treatment recommendations. Results
FIGURE I. Fine-needle aspiration biopsy instrumentation. Cameco IO-mL Syringe Pistol (Precision Dynamics, San Fernando, CA); IO-mL Luer-Lok syringe, intravenous extension tubing, and 23.gauge needle (Becton-Dickinson & Co, Rutherford, NJ); 23-gauge, S-cm aspiration biopsy needle ([mad, Grand Rapids, MI).
A total of 213 FNABs of head and neck lesions were performed on 209 patients at our institution between May 1987 and March 1990. Four patients underwent FNAB of lesions in two different locations. Of the 213 masses, 128 were located in nonspecific areas of the neck, 71 were located in the thyroid region, 10 in the parotid region, 6 on the
264
FINE-NEEDLE
scalp, and 4 within the oral cavity. Cytologic examination of the biopsy specimens resulted in 86 (40.3%) being diagnosed as malignant, 96 (45.1%) as benign, and 31 (14.6%) as indeterminant (Table 1). A specific cytologic diagnosis was made in 182 (85.5%) of the cases (Table 2). The two most commonly occurring lesions were squamous cell carcinoma (31 cases) and nodular goiter (30 cases). Thirty-one (14.6%) of the lesions were reported as indeterminant because the amount of aspirate material obtained was insufftcient. When the specific cytologic diagnostic percentages were compared for each calendar year of the study, an overall pattern of improvement was noted. The diagnostic percentages for the years 1987 through 1990 were 76.5%, 83.1%, 86.7% and 92.7%, respectively (Fig 2). Of the 209 patients on whom FNAB was performed, 110 subsequently underwent a surgical procedure with histologic examination of the excised tissue. Comparison of the cytologic diagnoses obtained by FNAB with the surgical diagnoses is shown in Table 3. The FNAB cytologic diagnosis agreed with the surgical histologic diagnosis in 99 of 110 specimens, for an overall specific diagnostic rate of 90%. The false-positive rate associated with FNAB in this series was 0.5%, and was based on one aspirate from a mass in the submandibular region that was reported as poorly differentiated carcinoma consistent with metastatic squamous cell carcinoma, but which proved to be necrotizing sialometaplasia secondary to radiation therapy of the neck. The true false-negative rate (FN,) was 2.3%. Cases in which FNAB resulted in a false-negative cytologic diagnosis included two papillary carcinomas of the thyroid, one that was misdiagnosed as a nodular goiter and the other as benign thyroid tissue; one nonHodgkins lymphoma that was misdiagnosed as a lipoma; a carotid body tumor misdiagnosed as a salivary adenoma; and a follicular carcinoma of the thyroid gland misdiagnosed as benign thyroid tissue. Other lesions that were incorrectly diagnosed by FNAB included a squamous cell carcinoma that was reported as mucoepidermoid carcinoma; two Table 1. Region
Table 2.
BIOPSY OF THE HEAD AND NECK
Cytologic Diagnoses by FNAB 31 30 12 9 9 8 8 6 6 6 5 5 5 5 4 4 4 3 3 3 3 2 2
Squamous cell carcinoma Nodular goiter Adenocarcinoma Branchial cleft cyst Undifferentiated carcinoma Reactive hyperplasia (lymph node) Non-small cell carcinoma Thyroglossal duct cyst Thyroid cyst (not otherwise specified) Malignant lymphoma Sialadenitis Pleomorphic adenoma Follicular lesion Papillary carcinoma Melanoma Mucoepidermoid carcinoma Warthin’s tumor Lipoma Small cell carcinoma Hurthle cell adenoma Lymphocytic thyroiditis Actinomycosis Lymphoepithelial lesion Myxoma Multiple myeloma Abcess Myelogenous leukemia Adenocystic carcinoma Folliculitis Scrofula Cystic hygroma Carotid body tumor
1 1 1
1 1 1 1
I 1
Total
182
follicular adenomas of the thyroid misdiagnosed as a nodular goiter and a Hurthle cell adenoma; and two nodular goiters misdiagnosed as a follicular lesion and a Hurthle cell adenoma. The sensitivity of FNAB in determining a malignant diagnosis was found to be 94.5% when FN, was used, 70.5% with FN,, and 81.1% with FN,. Calculated on a yearly basis (using FN,), sensitivity improved from a low of 57% in 1987 to a high of 91.7% in 1990 (Fig 2). Specificity was calculated at 99%. The negative predictive value of FNAB in determining a benign diagnosis was calculated as 95.1% using FN,, 72.7% using FN,, and 82.8% us-
Biopsy Site and General Diagnostic Category Total
Malignant
Benign
Indeterminant
FalsePositive
FalseNegative
Thyroid Neck Parotid Scalp Oral
71 128 10 6 4
11 65 4 5 1
45 45 4 1 1
13 14 2 0 2
0 1 0 0 0
2 3 0 0 0
Total (%)
213 (100)
86 (40.3)
96 (45.1)
31 (14.6)
1 (0.5)
5 (2.3)
SCHELKUN
AND GRUNDY
265
100%
-
90% -
80% 70% FIGURE 2. Yearly experience with fine-needle aspiration biopsy.
Specific cytologic diagnosis
60% -
: Sensitivity l
Negative predictive value
50% -
1987 ing FN,. When this figure was determined for each calendar year (using FN,), marked improvement (from 70% to 88.9%) was again noted with increasing experience. The positive predictive value for malignant diagnosis was 98.9%. Discussion A localized mass in the head and neck region is generally indicative of trauma, an inflammatory or infectious process, a benign or malignant neoplasm, or a cyst. The patient’s history and clinical findings will usually provide differential diagnostic information. However, in most cases morphologic examination is necessary to arrive at a definitive diagnosis and determine whether observation, medical therapy, chemotherapy, radiotherapy, or an operative procedure is indicated. Open surgical biopsy, especially of less superficial lesions of the head and neck, usually requires a general anesthetic. In some cases, such as primary or metastatic cancers, open biopsy may even be contraindicated due to the risk of spreading tumor cells through the planes of an open incisional or excisional biopsy.5*‘0*‘2 Scattered tumor cells and tissues in the area of open biopsy are also rendered relatively anoxic, making sterilization of the area by radiation treatment more difficult. I3 Fine-needle as-
piration biopsy, on the other hand, requires no more than a local anesthetic and, in experienced hands, can be performed in about 15 minutes. In inflammatory, traumatic or other reactive lesions, a benign FNAB allows observation, medical treatment and follow-up until resolution, thus avoiding an unnecessary surgical procedure. The complication rate associated with FNAB has been reported as zero by a number of authors.9*‘4~‘s In our hands, complications have been limited to a small hematoma, an area of ecchymosis, and some tenderness of the biopsy site. Previous authors have reviewed the literature for evidence of tumor seeding by FNAB,9q’2*‘5*‘6and although there have been rare reports in association with pancreatic and lung tumors, this has not been reported for FNAB of head and neck masses. Engzell and coworkers experimentally tried to produce tumor seeding either locally or distantly through FNAB in laboratory animals and failed. ” Qizilbash tattooed the FNAB needle tracks of tumors and then sectioned the needle tracks, but found no evidence of local tumor implantation. ‘* Our investigation of 213 FNAB of head and neck masses yielded a false-positive rate of 0.5% and a false-negative rate of 2.3%. The one false-positive lesion in our series was a submandibular mass in a patient who had undergone a full course of radiation
266 Table 3.
FINE-NEEDLE
Comparison of Surgical Histologic Diagnosis With FNAB Cytologic Diagnosis
Surgical Histologic Diagnosis Papillary carcinoma Benign thyroid cyst Follicular adenoma Follicular carcinoma Nodular goiter Lymphocytic thyroiditis Squamous cell carcinoma Adenocarcinoma Non-small cell carcinoma Melanoma Lymphoma Undifferentiated carcinoma SmalI cell carcinoma Myelogenous leukemia Carotid body tumor Pleomorphic adenoma Warthin’s tumor Mucoepidermoid carcinoma Sialadenitis Adenocystic carcinoma Necrotizing sialometaplasia Branchial cleft cyst Thyroglossal duct cyst Lipoma Actinomycosis Scrofula Reactive hyperplasia (lymph node) Totals (%)
BIOPSY OF THE HEAD AND NECK
Total Surgical Cases I 2 7 1 4 1 21 9 6 4 6 4 3 1 2 3 3 4 3 1 1 5 3 2 2 1 4 110 (100)
treatment for a recurrent squamous cell carcinoma of the tongue. The aspirated cells were atypical and malignant-appearing, and a diagnosis of poorly differentiated metastatic squamous cell carcinoma was made. The histology of the surgical specimens, however, revealed necrotizing sialometaplasia of the submandibular gland secondary to radiation therapy. Qizilbash and Young recommend a high degree of caution in the interpretation of salivary gland lesions in patients who have received radiation therapy,5 and suggest that radiation treatment of any area where FNAB is performed will make cytologic diagnosis difficult. The false-negatives found in this study were almost evenly divided between lesions located in the thyroid region (3) and lesions in nonspecific areas of the neck (2). Two of the thyroid false negatives occurred early in our experience, prior to using the modified FNAB technique previously described. In the third case, FNAB was performed by the referring surgeon, and on review was found to be a lessthan-adequate aspiration. The neck mass falsenegatives included a malignant lymphoma misdiagnosed as a lipoma, suggesting that fat was aspirated
Concordant FNAB Diagnosis 5 2 5 0 2 I 20 9 6 4 5 4 3 1 1 3 3 4 3 1 0 5 3 2 2 1 4 99 (90)
Discordant FNAB Diagnosis 2 (nodular goiter, benign thyroid) 0 2 (nodular goiter, Hurthle cell adenoma) 1 (benign thyroid) 2 (follicuiar lesion, Hurthle cell adenoma) 0 1 (mucoepidermoid carcinoma) 0 0 0 1 (lipoma) 0 0 0 1 (adenoma) 0 0 0 0 0 1 (undiierentiated carcinoma) 0 0 0 0 0 0 11 (10)
and not the mass. The carotid body tumor misdiagnosed as a salivary adenoma was also encountered early in our experience. A second carotid body tumor encountered at a later date was correctly diagnosed by FNAB. Our ability to render a cytologic diagnosis in 85.5% of the 213 FNAB cases compares well with other published reports.2”~6-7,9,12,‘4,15Improvement in this figure from 76.5% to 92.7% over the 4 calendar years of the study suggests that our diagnostic capabilities using FNAB have improved with experience . Statistical analysis of the sensitivity, specificity, and predictive value of FNAB confirms the basic premise that FNAB is a useful diagnostic tool. Sensitivity, which addresses the question “If a person has disease, how likely is it that a particular test for that disease is positive?,“” was found to be 81.1% for the total series. Specificity, which addresses the question “If a person does not have disease, how likely is the test to be negative?,” was 99%. These figures compare well with other reports.‘“15 Positive and negative predictive values for the total series were found to be 98.9% and 82.8%, respec-
SCHELKUN
267
AND GRUNDY
tively. Some element of bias was, of course, incorporated into our calculation of predictive values, given that our group of patients was not a reflection of the general population. There is, however, some information to be gained from this analysis and, therefore, it was included. In 90% of the 110 cases in which an operative procedure was later performed, the FNAB cytologic diagnosis concurred with the histologic diagnosis of the surgical specimen. This finding suggests that an accurate cytologic diagnosis can indeed be made using FNAB, paving the way for the development of an effective treatment plan. Our true specific diagnostic rate is probably greater than 90% because a large number of patients had benign or inflammatory lesions that required no operative procedure and resolved with medical therapy. Others required only a cytologic diagnosis for chemotherapy or radiotherapy. Our findings from this review lead us to conclude that FNAB is a sensitive and specific modality that assists in the diagnosis and management of head and neck masses. It is a safe procedure, virtually without risk to the patient, and patient acceptance is high. With training and experience, the procedure can be carried out in less than 15 minutes, and the results should be available within a few hours. It should be kept in mind, however, that FNAB is just one element in the diagnostic workup of a mass. With a clinically suspicious mass, a negative or nondiagnostic aspiration indicates only that a repeat aspiration may be necessary or that some other biopsy procedure may be required. The accuracy of diagnosis will vary not only with the experience of the pathologist, but also with the experience of the individual performing the biopsy and the quality of slide preparation. Finally, it is essential that there is full communication and understanding between the clinicians and pathologist if one is to make the best use of this diagnostic modality.
References 1. Martin HE, Ellis EB: Biopsy by needle puncture and aspiration. Ann Surg 92: 169, 1936 2. Eneroth CM. Franzen S. Zaiicek J: Cvtoloaic diaanosis on aspirate of 1000 salivary &nd tumors. &ta Ot&ryngol 224:168, 1967 3. Mavec P, Eneroth CM, Franzen S, et al: Aspiration biopsy of salivary gland tumors. Acta Otolaryngol 58471, 1964 4. Zajicek J: Aspiration Biopsy Cytology: I. Cytology of the Superdiaphragmatic Organs, vol 4. New York, NY, Karger, 1974 5. Qizilbash AH, Young JEM: Guide to Clinical Aspiration Biopsy: Head and Neck. New York, NY, Igaku-Shoin, 1988 6. Persson PS, Zettergren L: Cytologic diagnosis of salivary gland tumors by aspiration biopsy. Acta Cytol 17:351, 1973 7. Thompson J, Andreassen JC, Bangsbo C: Fine needle aspiration biopsy of tumors of the head and neck. J Laryngol Otol 86:1211, 1973 8. Meyers DS, Templer J, Davis WE, et al: Aspiration cytology for diagnosis of head and neck masses. Otolaryngol Head Neck Surg 86:650, 1978 9. Sismonis A, Merriam JM, Yamaguchi KT, et al: Diagnostic value of fine-needle aspiration biopsy in neoplasms of the head and neck. Otolaryngol Head Neck Surg 89:62, 1981 IO. Frable WJ: Thin Needle Aspiration Biopsy. Philadelphia, PA, Saunders, 1983 11. Feldman PS, Cove11JL, Kandos TF: Fine Needle Aspiration Cytology: Lymph Node, Thyroid and Salivary Gland. Chicago, IL, American Society for Clinical Pathology (ASCP) Press, 1989 12. Frable WJ: Thin-needle aspiration biopsy. Am J Clin Path01 65: 168, 1976 13. Fletcher GH: Textbook of Radiotherapy (ed 3). Philadelphia, PA, Lea & Febiger, 1980 14. Peters BR, Schnadig VJ, Quinn FB, et al: Interobserver variability in the interpretation of tine-needle aspiration biopsy of head and neck masses. Arch Otolaryngol Head Neck Surg 115:1438, 1989 15. Layfield CJ, Tan P, Glasgow BJ: Fine-needle aspiration of salivary gland lesions. Arch Pathol Lab Med II 1:346, 1987 16. Berg JW, Robbins GF: A late look at the safety of aspiration biopsy. Cancer 15:826, I%2 17. Engzell V, Esposti DL, Rubio C, et al: investigation on tumor spread in connection with aspiration biopsy. Acta Radio1 10:385, 1971 18. Qizilbash AH, Sianos J, Young JEM, et al: Fine needle aspiration cytology of major salivary glands. Acta Cytol 29:503, 1985 19. lngelfinger JA, Mosteller F, Thibodeau LA, et al: Biostatistics in Clinical Medicine. New York, NY. MacMillan, 1987
Surg
1991
Fine-Needle Aspiration Biopsy of Head and Neck Lesions P. MICHAEL SCHELKUN,
DDS, MD,* AND WALTER G. GRUNDY, MDt
The case records of 213 fine-needle aspiration biopsies (FNAB) of head and neck masses performed on 209 patients over a 3-year period were reviewed to assess the diagnostic accuracy and safety of this technique in comparison with surgical histologic examination. Cytologic diagnoses based on FNAB were compared with histologic diagnoses in 110 patients who underwent surgery. Based on cytology alone, 40.3% of the lesions were reported as malignant, 45.1% as benign, and 14.6% as indeterminant. A specific cytologic diagnosis was made in 85.5% of the cases. Cytologic diagnoses concurred with surgical histologic diagnoses 90% of the time. Fine-needle aspiration biopsy was found to have a false-positive rate of 0.5% and a false-negative rate of 2.3%. The sensitivity and specificity of FNAB in determining a malignant diagnosis were 81 .l% and 99%, respectively. Positive and negative predictive values were calculated at 98.9% and 82.8%, respectively. Diagnostic rate, sensitivity, and negative predictive value increased consistently throughout the study period, indicating that the diagnostic accuracy of FNAB improved with experience.
and safety in comparison with the widely accepted standard of histologic examination of surgical specimens.
Although 60 years have passed since Martin and Ellis published their now famous treatise on fineneedle aspiration biopsy (FNAB),’ the American medical community has been slow in recognizing the value of this diagnostic technique. Over the past 30 years, FNAB has been widely applied on the European continent, particularly in Scandinavia, and has proved to be a safe, rapid, inexpensive, and accurate method for establishing a diagnosis and defining treatment.2-6 Based on the published European experience and recent experience at some American centers,‘-‘* it now appears that FNAB is gaining wider acceptance among clinicians. In a retrospective review, we analyzed our experience with FNAB of head and neck masses over the past 3 years to assess its diagnostic accuracy
Materials and Methods The case records of all patients who underwent FNAB of head and neck masses at our institution between May 1987 and March 1990 were reviewed. This included biopsies of all lesions located superior to the clavicles. Over 96% of the masses were biopsied and interpreted by one of the authors (WGG). NEEDLE
In all cases, the technique used to collect the biopsy specimens was based on one previously described by Zajicek and Frable,4,10 with some minor modifications. If the lesion could be easily palpated and stabilized, a lo-mL Luer-Lok (BectonDickinson, Rutherford, NJ) syringe in a Cameco Syringe Pistol Holder (Precision Dynamics, San Fernando, CA) with a 23-gauge needle was used. If the mass was small and/or relatively superficial, however, a 23-gauge needle was attached to one
* Resident,
Department of Surgery, Rush-Presbyterian-St Luke’s Medical Center, Chicago. t Clinical Assistant Professor, Department of Pathology, The University of Alabama in Huntsville School of Primary Medical Care, and Huntsville Hospital, Huntsville, AL. Address correspondence and reprint requests to Dr Schelkun: 3714 N Wayne St, Chicago, IL 60613. 0 1991 American geons
Association
of Oral and Maxillofacial
BIOPSY TECHNIQUE
Sur-
0278-2391/91/4903-0008$3.00/0
262
SCHELKUN
AND GRUNDY
end of an intravenous extension tube and the IO-mL Luer-Lok syringe to the other (Fig 1). An assistant was required to create negative pressure within the tubing apparatus. Following anesthetization of the skin with 1% lidocaine, two or three fine-needle aspirations were generally performed at each biopsy site. After each aspiration, a drop of material was placed on a slide and a smear was prepared. Approximately half of the slides were fixed in Reagent Alcohol (Fisher Scientific, Fairlawn, NJ), and half were air-dried. The needle and syringe were then rinsed with Muccolex (Lerner Laboratories, Pittsburgh, PA), a cytologic fixative. If the material was scant, the majority of slides were air-dried. The slides were routinely stained with Diff-Quik (American Scientific Products, McGaw Park, IL), Papanicolaou, and hematoxylin-eosin stains. Special stains were used when indicated. The solution of Muccolex and the aspirate was passed through a millipore filter and placed on slides and stained as previously described. Experience dictated a modification of this biopsy technique for thyroid gland aspirates. Because of the highly vascular nature of the thyroid gland, and the propensity to aspirate mostly blood, the use of a 23-gauge or 25gauge, 5-cm Inrad (Grand Rapids, MI) aspiration biopsy needle was adopted. We found that by using the natural capillary action of the needle, a more diagnostic aspirate could be obtained. STATISTICAL METHODS
Data obtained from the review of case records were first categorized according to the regional lo-
263
cation of the biopsy site: thyroid, neck, parotid, scalp, or intraoral. We then noted both the general cytologic diagnosis (malignant, benign, or indeterminant) and the specific cytologic diagnosis, when available, for each of the lesions. Cytologic findings were compared with the results of histological examination of surgical specimens in those patients who subsequently underwent an operative procedure to determine the specific diagnostic rate and number of false-positive and false-negative diagnoses obtained by FNAB. The diagnostic rate was calculated by dividing the number of cases in which there was agreement between the cytology and the surgical histology by the total number of cases that were treated surgically. The sensitivity and specificity of FNAB for determining malignancy, as well as positive and negative predictive values, were also calculated. Sensitivity was calculated as TP/(TP + FN), for which TP equals the number of lesions identified as malignant and FN is the number of false-negatives. Diagnostic specificity was calculated as TN/(TN + FP). for which TN equaIs the number of lesions identified as benign and FP equals the number of false-positives. Positive predictive value was calculated as TP/ (TP + FP), for which TP equals the number of lesions diagnosed as malignant and FP equals the number of false-positives. Negative predictive value was calculated as TN&TN + FN), for which TN equals the number of lesions identified as benign and FN equals the number of false-negatives. In our analysis, the false-negative values used in the calculation of sensitivity and negative predictive value were determined in three different ways. The first value (FN i), or true false-negative, included ah lesions with which cytologic diagnosis was designated as malignant or suspicious for malignancy and with benign surgical histology. The second value (FN,) included the true false-negatives plus the total number of indeterminants. The third value (FN3) was defined as the true false-negatives plus the indeterminants, minus the total number of patients within the indeterminant group who were noncompliant with treatment recommendations. Results
FIGURE I. Fine-needle aspiration biopsy instrumentation. Cameco IO-mL Syringe Pistol (Precision Dynamics, San Fernando, CA); IO-mL Luer-Lok syringe, intravenous extension tubing, and 23.gauge needle (Becton-Dickinson & Co, Rutherford, NJ); 23-gauge, S-cm aspiration biopsy needle ([mad, Grand Rapids, MI).
A total of 213 FNABs of head and neck lesions were performed on 209 patients at our institution between May 1987 and March 1990. Four patients underwent FNAB of lesions in two different locations. Of the 213 masses, 128 were located in nonspecific areas of the neck, 71 were located in the thyroid region, 10 in the parotid region, 6 on the
264
FINE-NEEDLE
scalp, and 4 within the oral cavity. Cytologic examination of the biopsy specimens resulted in 86 (40.3%) being diagnosed as malignant, 96 (45.1%) as benign, and 31 (14.6%) as indeterminant (Table 1). A specific cytologic diagnosis was made in 182 (85.5%) of the cases (Table 2). The two most commonly occurring lesions were squamous cell carcinoma (31 cases) and nodular goiter (30 cases). Thirty-one (14.6%) of the lesions were reported as indeterminant because the amount of aspirate material obtained was insufftcient. When the specific cytologic diagnostic percentages were compared for each calendar year of the study, an overall pattern of improvement was noted. The diagnostic percentages for the years 1987 through 1990 were 76.5%, 83.1%, 86.7% and 92.7%, respectively (Fig 2). Of the 209 patients on whom FNAB was performed, 110 subsequently underwent a surgical procedure with histologic examination of the excised tissue. Comparison of the cytologic diagnoses obtained by FNAB with the surgical diagnoses is shown in Table 3. The FNAB cytologic diagnosis agreed with the surgical histologic diagnosis in 99 of 110 specimens, for an overall specific diagnostic rate of 90%. The false-positive rate associated with FNAB in this series was 0.5%, and was based on one aspirate from a mass in the submandibular region that was reported as poorly differentiated carcinoma consistent with metastatic squamous cell carcinoma, but which proved to be necrotizing sialometaplasia secondary to radiation therapy of the neck. The true false-negative rate (FN,) was 2.3%. Cases in which FNAB resulted in a false-negative cytologic diagnosis included two papillary carcinomas of the thyroid, one that was misdiagnosed as a nodular goiter and the other as benign thyroid tissue; one nonHodgkins lymphoma that was misdiagnosed as a lipoma; a carotid body tumor misdiagnosed as a salivary adenoma; and a follicular carcinoma of the thyroid gland misdiagnosed as benign thyroid tissue. Other lesions that were incorrectly diagnosed by FNAB included a squamous cell carcinoma that was reported as mucoepidermoid carcinoma; two Table 1. Region
Table 2.
BIOPSY OF THE HEAD AND NECK
Cytologic Diagnoses by FNAB 31 30 12 9 9 8 8 6 6 6 5 5 5 5 4 4 4 3 3 3 3 2 2
Squamous cell carcinoma Nodular goiter Adenocarcinoma Branchial cleft cyst Undifferentiated carcinoma Reactive hyperplasia (lymph node) Non-small cell carcinoma Thyroglossal duct cyst Thyroid cyst (not otherwise specified) Malignant lymphoma Sialadenitis Pleomorphic adenoma Follicular lesion Papillary carcinoma Melanoma Mucoepidermoid carcinoma Warthin’s tumor Lipoma Small cell carcinoma Hurthle cell adenoma Lymphocytic thyroiditis Actinomycosis Lymphoepithelial lesion Myxoma Multiple myeloma Abcess Myelogenous leukemia Adenocystic carcinoma Folliculitis Scrofula Cystic hygroma Carotid body tumor
1 1 1
1 1 1 1
I 1
Total
182
follicular adenomas of the thyroid misdiagnosed as a nodular goiter and a Hurthle cell adenoma; and two nodular goiters misdiagnosed as a follicular lesion and a Hurthle cell adenoma. The sensitivity of FNAB in determining a malignant diagnosis was found to be 94.5% when FN, was used, 70.5% with FN,, and 81.1% with FN,. Calculated on a yearly basis (using FN,), sensitivity improved from a low of 57% in 1987 to a high of 91.7% in 1990 (Fig 2). Specificity was calculated at 99%. The negative predictive value of FNAB in determining a benign diagnosis was calculated as 95.1% using FN,, 72.7% using FN,, and 82.8% us-
Biopsy Site and General Diagnostic Category Total
Malignant
Benign
Indeterminant
FalsePositive
FalseNegative
Thyroid Neck Parotid Scalp Oral
71 128 10 6 4
11 65 4 5 1
45 45 4 1 1
13 14 2 0 2
0 1 0 0 0
2 3 0 0 0
Total (%)
213 (100)
86 (40.3)
96 (45.1)
31 (14.6)
1 (0.5)
5 (2.3)
SCHELKUN
AND GRUNDY
265
100%
-
90% -
80% 70% FIGURE 2. Yearly experience with fine-needle aspiration biopsy.
Specific cytologic diagnosis
60% -
: Sensitivity l
Negative predictive value
50% -
1987 ing FN,. When this figure was determined for each calendar year (using FN,), marked improvement (from 70% to 88.9%) was again noted with increasing experience. The positive predictive value for malignant diagnosis was 98.9%. Discussion A localized mass in the head and neck region is generally indicative of trauma, an inflammatory or infectious process, a benign or malignant neoplasm, or a cyst. The patient’s history and clinical findings will usually provide differential diagnostic information. However, in most cases morphologic examination is necessary to arrive at a definitive diagnosis and determine whether observation, medical therapy, chemotherapy, radiotherapy, or an operative procedure is indicated. Open surgical biopsy, especially of less superficial lesions of the head and neck, usually requires a general anesthetic. In some cases, such as primary or metastatic cancers, open biopsy may even be contraindicated due to the risk of spreading tumor cells through the planes of an open incisional or excisional biopsy.5*‘0*‘2 Scattered tumor cells and tissues in the area of open biopsy are also rendered relatively anoxic, making sterilization of the area by radiation treatment more difficult. I3 Fine-needle as-
piration biopsy, on the other hand, requires no more than a local anesthetic and, in experienced hands, can be performed in about 15 minutes. In inflammatory, traumatic or other reactive lesions, a benign FNAB allows observation, medical treatment and follow-up until resolution, thus avoiding an unnecessary surgical procedure. The complication rate associated with FNAB has been reported as zero by a number of authors.9*‘4~‘s In our hands, complications have been limited to a small hematoma, an area of ecchymosis, and some tenderness of the biopsy site. Previous authors have reviewed the literature for evidence of tumor seeding by FNAB,9q’2*‘5*‘6and although there have been rare reports in association with pancreatic and lung tumors, this has not been reported for FNAB of head and neck masses. Engzell and coworkers experimentally tried to produce tumor seeding either locally or distantly through FNAB in laboratory animals and failed. ” Qizilbash tattooed the FNAB needle tracks of tumors and then sectioned the needle tracks, but found no evidence of local tumor implantation. ‘* Our investigation of 213 FNAB of head and neck masses yielded a false-positive rate of 0.5% and a false-negative rate of 2.3%. The one false-positive lesion in our series was a submandibular mass in a patient who had undergone a full course of radiation
266 Table 3.
FINE-NEEDLE
Comparison of Surgical Histologic Diagnosis With FNAB Cytologic Diagnosis
Surgical Histologic Diagnosis Papillary carcinoma Benign thyroid cyst Follicular adenoma Follicular carcinoma Nodular goiter Lymphocytic thyroiditis Squamous cell carcinoma Adenocarcinoma Non-small cell carcinoma Melanoma Lymphoma Undifferentiated carcinoma SmalI cell carcinoma Myelogenous leukemia Carotid body tumor Pleomorphic adenoma Warthin’s tumor Mucoepidermoid carcinoma Sialadenitis Adenocystic carcinoma Necrotizing sialometaplasia Branchial cleft cyst Thyroglossal duct cyst Lipoma Actinomycosis Scrofula Reactive hyperplasia (lymph node) Totals (%)
BIOPSY OF THE HEAD AND NECK
Total Surgical Cases I 2 7 1 4 1 21 9 6 4 6 4 3 1 2 3 3 4 3 1 1 5 3 2 2 1 4 110 (100)
treatment for a recurrent squamous cell carcinoma of the tongue. The aspirated cells were atypical and malignant-appearing, and a diagnosis of poorly differentiated metastatic squamous cell carcinoma was made. The histology of the surgical specimens, however, revealed necrotizing sialometaplasia of the submandibular gland secondary to radiation therapy. Qizilbash and Young recommend a high degree of caution in the interpretation of salivary gland lesions in patients who have received radiation therapy,5 and suggest that radiation treatment of any area where FNAB is performed will make cytologic diagnosis difficult. The false-negatives found in this study were almost evenly divided between lesions located in the thyroid region (3) and lesions in nonspecific areas of the neck (2). Two of the thyroid false negatives occurred early in our experience, prior to using the modified FNAB technique previously described. In the third case, FNAB was performed by the referring surgeon, and on review was found to be a lessthan-adequate aspiration. The neck mass falsenegatives included a malignant lymphoma misdiagnosed as a lipoma, suggesting that fat was aspirated
Concordant FNAB Diagnosis 5 2 5 0 2 I 20 9 6 4 5 4 3 1 1 3 3 4 3 1 0 5 3 2 2 1 4 99 (90)
Discordant FNAB Diagnosis 2 (nodular goiter, benign thyroid) 0 2 (nodular goiter, Hurthle cell adenoma) 1 (benign thyroid) 2 (follicuiar lesion, Hurthle cell adenoma) 0 1 (mucoepidermoid carcinoma) 0 0 0 1 (lipoma) 0 0 0 1 (adenoma) 0 0 0 0 0 1 (undiierentiated carcinoma) 0 0 0 0 0 0 11 (10)
and not the mass. The carotid body tumor misdiagnosed as a salivary adenoma was also encountered early in our experience. A second carotid body tumor encountered at a later date was correctly diagnosed by FNAB. Our ability to render a cytologic diagnosis in 85.5% of the 213 FNAB cases compares well with other published reports.2”~6-7,9,12,‘4,15Improvement in this figure from 76.5% to 92.7% over the 4 calendar years of the study suggests that our diagnostic capabilities using FNAB have improved with experience . Statistical analysis of the sensitivity, specificity, and predictive value of FNAB confirms the basic premise that FNAB is a useful diagnostic tool. Sensitivity, which addresses the question “If a person has disease, how likely is it that a particular test for that disease is positive?,“” was found to be 81.1% for the total series. Specificity, which addresses the question “If a person does not have disease, how likely is the test to be negative?,” was 99%. These figures compare well with other reports.‘“15 Positive and negative predictive values for the total series were found to be 98.9% and 82.8%, respec-
SCHELKUN
267
AND GRUNDY
tively. Some element of bias was, of course, incorporated into our calculation of predictive values, given that our group of patients was not a reflection of the general population. There is, however, some information to be gained from this analysis and, therefore, it was included. In 90% of the 110 cases in which an operative procedure was later performed, the FNAB cytologic diagnosis concurred with the histologic diagnosis of the surgical specimen. This finding suggests that an accurate cytologic diagnosis can indeed be made using FNAB, paving the way for the development of an effective treatment plan. Our true specific diagnostic rate is probably greater than 90% because a large number of patients had benign or inflammatory lesions that required no operative procedure and resolved with medical therapy. Others required only a cytologic diagnosis for chemotherapy or radiotherapy. Our findings from this review lead us to conclude that FNAB is a sensitive and specific modality that assists in the diagnosis and management of head and neck masses. It is a safe procedure, virtually without risk to the patient, and patient acceptance is high. With training and experience, the procedure can be carried out in less than 15 minutes, and the results should be available within a few hours. It should be kept in mind, however, that FNAB is just one element in the diagnostic workup of a mass. With a clinically suspicious mass, a negative or nondiagnostic aspiration indicates only that a repeat aspiration may be necessary or that some other biopsy procedure may be required. The accuracy of diagnosis will vary not only with the experience of the pathologist, but also with the experience of the individual performing the biopsy and the quality of slide preparation. Finally, it is essential that there is full communication and understanding between the clinicians and pathologist if one is to make the best use of this diagnostic modality.
References 1. Martin HE, Ellis EB: Biopsy by needle puncture and aspiration. Ann Surg 92: 169, 1936 2. Eneroth CM. Franzen S. Zaiicek J: Cvtoloaic diaanosis on aspirate of 1000 salivary &nd tumors. &ta Ot&ryngol 224:168, 1967 3. Mavec P, Eneroth CM, Franzen S, et al: Aspiration biopsy of salivary gland tumors. Acta Otolaryngol 58471, 1964 4. Zajicek J: Aspiration Biopsy Cytology: I. Cytology of the Superdiaphragmatic Organs, vol 4. New York, NY, Karger, 1974 5. Qizilbash AH, Young JEM: Guide to Clinical Aspiration Biopsy: Head and Neck. New York, NY, Igaku-Shoin, 1988 6. Persson PS, Zettergren L: Cytologic diagnosis of salivary gland tumors by aspiration biopsy. Acta Cytol 17:351, 1973 7. Thompson J, Andreassen JC, Bangsbo C: Fine needle aspiration biopsy of tumors of the head and neck. J Laryngol Otol 86:1211, 1973 8. Meyers DS, Templer J, Davis WE, et al: Aspiration cytology for diagnosis of head and neck masses. Otolaryngol Head Neck Surg 86:650, 1978 9. Sismonis A, Merriam JM, Yamaguchi KT, et al: Diagnostic value of fine-needle aspiration biopsy in neoplasms of the head and neck. Otolaryngol Head Neck Surg 89:62, 1981 IO. Frable WJ: Thin Needle Aspiration Biopsy. Philadelphia, PA, Saunders, 1983 11. Feldman PS, Cove11JL, Kandos TF: Fine Needle Aspiration Cytology: Lymph Node, Thyroid and Salivary Gland. Chicago, IL, American Society for Clinical Pathology (ASCP) Press, 1989 12. Frable WJ: Thin-needle aspiration biopsy. Am J Clin Path01 65: 168, 1976 13. Fletcher GH: Textbook of Radiotherapy (ed 3). Philadelphia, PA, Lea & Febiger, 1980 14. Peters BR, Schnadig VJ, Quinn FB, et al: Interobserver variability in the interpretation of tine-needle aspiration biopsy of head and neck masses. Arch Otolaryngol Head Neck Surg 115:1438, 1989 15. Layfield CJ, Tan P, Glasgow BJ: Fine-needle aspiration of salivary gland lesions. Arch Pathol Lab Med II 1:346, 1987 16. Berg JW, Robbins GF: A late look at the safety of aspiration biopsy. Cancer 15:826, I%2 17. Engzell V, Esposti DL, Rubio C, et al: investigation on tumor spread in connection with aspiration biopsy. Acta Radio1 10:385, 1971 18. Qizilbash AH, Sianos J, Young JEM, et al: Fine needle aspiration cytology of major salivary glands. Acta Cytol 29:503, 1985 19. lngelfinger JA, Mosteller F, Thibodeau LA, et al: Biostatistics in Clinical Medicine. New York, NY. MacMillan, 1987
