Diagnostic accuracy of palpation-guided and image-guided fine-needle aspiration biopsy of the thyroid.

Abstract

We conducted a retrospective study to compare the sensitivity and specificity of traditional palpation-guided fine-needle aspiration biopsy (FNAB) performed by clinicians and pathologists with that of image-guided FNAB performed by radiologists for the evaluation of thyroid nodules. We reviewed the medical records of 89 patients' who had undergone thyroid FNAB and subsequent surgical excision and pathology. Of this" group, 58 patients had undergone palpation-guided FNAB performed by a clinician, 20 had undergone palpation-guided FNAB performed by a pathologist, and 11 had undergone image-guided FNAB performed by a radiologist. The sensitivity of the three techniques was 86, 100, and 100%, respectively, and the sepecificity was 78, 94, and 44%; there were no statistically significant differences in sensitivity or specificity among the three groups. Our data indicate that FNAB of the thyroid can be performed with equal reliability by clinicians, pathologists, and radiologists.

Introduction

The reported incidence of clinically evident thyroid nodules in North America ranges from 4 to 7%. (1) The malignancy rate of palpable thyroid nodules is less than 5%. (2,3) The evaluation of these nodules includes radionuclide scanning, ultrasonography, and fine-needle aspiration biopsy (FNAB). FNAB has become accepted as a reliable tool for the investigation of thyroid nodules.

The sensitivity of palpation-guided thyroid FNABs has been reported to range from 65 to 98%, with specificity ranging from 72 to 100%. (4-8) Over the past decade, there has been an increasing trend toward the use of image-guided FNAB for the evaluation of head and neck lesions, including thyroid nodules. (9-12) Ultrasonographic guidance is the most popular imaging modality for this purpose, but computed tomography (CT) has been used, as well. The sensitivity (60 to 100%) and specificity (87 to 100%) of image-guided thyroid FNAB are also relatively high, and this technique yields a high rate of adequate samples (96%). (1,6,8,9) Image guidance is increasingly being used in patients who have small or nonpalpable lesions, incidentalomas, and lesions in difficult-to-access locations, as well as in patients whose previous FNABs were nondiagnostic. (13-15)

In view of the increasing use of image-guided FNAB performed by radiologists, we set out to determine whether the results they obtain are significantly different from the results obtained by palpation-guided FNAB performed by clinicians and pathologists.

Patients and methods

We retrospectively reviewed the medical records of 484 patients who had undergone FNAB of a thyroid nodule between Aug. 1, 1993, and Aug. 31, 1999, at Norton Hospital, an affiliate of the University of Louisville (Ky.) School of Medicine. The vast majority of these FNABs (89%) had been palpation-guided.

By cross-referencing pathology records, we learned that 89 of these patients had undergone subsequent surgical excision (thyroid lobectomy) and pathologic examination. This group was made up of 36 men and 53 women, aged 19 to 77 years (mean: 55.4). The mean nodule size was 2.8 cm. The records of these patients allowed us to compare their preoperative FNAB results with the subsequent pathologic findings.

The FNAB techniques used at our institution were similar to those used in other studies. (2,5) Clinicians evaluated their own samples, which were submitted in the appropriate fixative solution. FNAB aspirates obtained by radiologists and pathologists were assessed for adequacy of content at the time of aspiration. Additional samples were obtained when specimens did not contain satisfactory material for analysis.

We then performed a data analysis to calculate the sensitivity and specificity of palpation-guided and image-guided FNABs. Sensitivity was calculated by dividing the number of true-positive (TP) results by the number of true positives plus the number of false-negative (FN) results--that is, sensitivity = TP/(TP + FN). Specificity was calculated by dividing the number of true-negative (TN) results by the number of true negatives plus the number of false-positive (FP) results--that is, specificity = TN/(TN + FP). The chi-square ([chi square]) test and analysis of variance (ANOVA) were used to compare the data. (16) The positive predictive value (PPV) was calculated by dividing the total number of true positives by the total number of all positive tests--that is, PPV = TP/(TP + FP). An FNAB result was considered to be positive if it was interpreted as either malignant or suspicious and negative if it was interpreted as benign. The inclusion of suspicious FNABs in the positive group was considered appropriate because the aspirations were performed in order to identify patients who were candidates for definitive histologic analysis (excisional biopsy-thyroid lobectomy). Nondiagnostic FNABs were excluded from the calculations.

Results

Of 89 patients, 58 (65%) had undergone palpation-guided FNAB performed by a clinician, 20 (22%) had undergone palpation-guided FNAB performed by a pathologist, and 11 (12%) had undergone image-guided FNAB performed by a radiologist (table).

Clinician group. Clinician-performed FNAB indicated that 26 of the 58 specimens (45%) were benign, 4 (7%) were malignant, 21 (36%) were suspicious, and 7 (12%) were nondiagnostic. Surgical pathology revealed that 46 (79%) were benign and 12 (21%) were malignant. The 12 malignancies included 7 papillary carcinomas, 3 Hurthle cell carcinomas, and 2 follicular carcinomas. Two malignancies were incidental findings. One case involved the detection of a microscopic focus of papillary carcinoma at a site away from the lesion sampled; the other tumor was a Hurthle cell carcinoma that also was located apart from the lesion sampled. The sensitivity of clinician-performed FNAB was 86%, the specificity was 78%, and the PPV was 48%.

Pathologist group. Of the 20 pathologist-performed FNABs, 12 (60%) were benign, 4 (20%) were malignant, 2 (10%) were suspicious, and 2 (10%) were nondiagnostic. On surgical pathology, 15 (75%) were benign and 5 (25%) were malignant. The malignancies included 4 papillary carcinomas and 1 medullary carcinoma. The sensitivity was 100%, the specificity was 94%, and the PPV was 83%.

Radiologist group. On image-guided FNAB, 4 of the 11 specimens (36%) were benign, 1 (9%) was malignant, and 6 (55%) were suspicious. There were no nondiagnostic FNABs in this group. Surgical pathology identified 9 as benign (82%) and 2 as malignant (18%). Both of the malignancies were papillary carcinomas. The sensitivity, specificity, and PPV were 100, 44, and 29%, respectively.

Statistical analysis. According to the [chi square] test and ANOVA, there were no statistically significant differences among the three groups with respect to sensitivity or specificity (p = 0.215).

Discussion

Traditionally, the work-up and diagnosis of thyroid nodules has been based on the results of palpation-guided FNAB. The goal of FNAB is to identify which patients with thyroid nodules are at increased risk of thyroid malignancy and therefore will require a subsequent surgical biopsy. Using this technique, the incidence of malignancy in surgical specimens has increased from 14 to 29%. (7) This higher yield of malignancy was accomplished by reducing the number of patients who are candidates for excisional biopsy (thyroid lobectomy). Patients with clearly benign FNAB pathology are not candidates for excisional biopsy and are placed on observation status. This reduced the total number of patients who undergo excisional biopsy.

While palpation-guided FNAB has long been the standard, image-guided FNAB with ultrasonography or CT has become increasingly popular. (9-12) The advantages of image-guided FNAB include its ability to evaluate lesions that are difficult to identify on palpation, including small lesions, lesions that have a large cystic component, deep or posterior masses, and substernal masses. It can also detect incidental nodules and provide additional information on nodules that previously yielded nondiagnostic material on FNAB. Image guidance also identifies blood vessels in the thyroid and otherwise aids in optimizing the localization of a biopsy site. Finally, it allows for verification of the needle position within the mass.

The sensitivities of the three types of FNAB assessed in out study--86% for palpation-guided FNAB by clinicians and 100% each for palpation-guided FNAB by pathologists and image-guided FNAB by radiologists--were not significantly different, and they correlate well with data in previously published reports. (9-12) Likewise, the [chi square] test and ANOVA did not indicate any statistically significant difference among the specificities of the three types of FNAB, which ranged from 44 to 94%.

The practice of classifying a suspicious FNAB finding as a positive result is controversial, and it results in a much higher specificity than would otherwise be the case. Yet the purpose of FNAB is not only to identify patients with clearly positive findings for malignancy, but to also identify those whose findings warrant a high index of suspicion for malignancy. Therefore, we consider the inclusion of suspicious results as positive findings to be clinically valid.

There were no nondiagnostic or inadequate specimens among our small sample of image-guided FNABs. We consider the low nondiagnostic or inadequate specimen rate for image-guided FNAB (0%) as the major advantage of this approach over the clinician and pathologist palpation-guided rates (12 and 10%, respectively). Based on this finding and on the high sensitivity of image-guided FNAB, we predict that an image-guided FNAB team approach, in which precise needle placement is performed by a radiologist or surgeon and the specimen's adequacy is immediately verified by a pathologist, will gradually replace palpation-guided FNAB in many clinical settings. The presence of a cytopathology technician or a pathologist during FNAB has been shown to increase the proportion of diagnostic samples. (4)

One arguable limitation of our study was the lack of a strict protocol. We believe that our study design represents a more accurate reflection of the normal clinical practice environment than a protocol conducted under rigid controls. The FNABs in this study were obtained by a number of clinicians, pathologists, and radiologists. While the techniques they employed were generally consistent, they were not tightly controlled in terms of the exact number of aspirations that were obtained from each patient. Similarly, there were no strict criteria to follow when determining the adequacy of a specimen. In each situation, they used their best judgment.

Obviously, pathologists have an advantage in determining whether a sample is adequate, while clinicians simply base their decision on a visual assessment of the specimen. We do not suggest that the clinicians' method is the best way of judging the adequacy of a sample; we only suggest that this method reflects widespread current practice. Under optimal circumstances, a specific and detailed protocol would explicitly define the technical methodology for obtaining and evaluating FNAB specimens, but this is not practical in everyday practice.

An unambiguous limitation of our study was the small number of patients in each of the three groups.

In conclusion, our data indicate that FNAB of the thyroid can be performed with equal reliability by a clinician, pathologist, or radiologist. We did not find any statistically significant differences among the three groups in terms of sensitivity and specificity. However, as image-guided FNAB becomes more popular and data on more patients become available, a statistically significant difference may yet emerge.

Table. Findings on fine-needle aspiration biopsy (FNAB) and
subsequent pathology of excisional biopsy in the three
groups (N = 89)

                             Clinician group        Pathologist group
                              n = 58 (65%)             n = 20 (22%)

                           FNAB       Pathology     FNAB    Pathology

Result (n [%])           26 (45)       46 (79)     12 (60)   15 (75)
  Benign                  4 (7)        12 (21)      4 (20)    5 (25)
  Malignant              21 (36)         --         2 (10)     --
  Suspicious              7 (12)         --         2 (10)     --
  Nondiagnostic

Sensitivity (%)            86                        100
Specificity (%)            78                         94
Positive predictive        48                         83
  value (%)

                          Radiologist group
                             n = 20 (22%)

                          FNAB     Pathology

Result (n [%])           4 (36)     9 (82)
  Benign                 1 (9)      2 (18)
  Malignant              6 (55)      --
  Suspicious             0 (0)       --
  Nondiagnostic

Sensitivity (%)          100
Specificity (%)           44
Positive predictive       29
  value (%)

References

(1.) Cochand-Priollet B, Guillausseau PJ, Chagnon S, et al. The diagnostic value of fine-needle aspiration biopsy under ultrasonography in nonfunctional thyroid nodules: A prospective study comparing cytologic and histologic findings. Am J Med 1994;97:152-7.

(2.) Gharib H, Goellner JR. Fine-needle aspiration biopsy of the thyroid: An appraisal. Ann Intern Med 1993;118:282-9.

(3.) Hundahl SA, Fleming ID, Fremgen AM, Menck HR. A National Cancer Data Base report on 53,856 cases of thyroid carcinoma treated in the U.S., 1985-1995. Cancer 1998;83:2638-48.

(4.) Akerman M, Tennvall J, Biorklund A, et al. Sensitivity and specificity of fine needle aspiration cytology in the diagnosis of tumors of the thyroid gland. Acta Cytol 1985;29:850-5.

(5.)Baloch ZW, Sack MJ, Yu GH, et al. Fine-needle aspiration of thyroid: An institutional experience. Thyroid 1998;8:565-9.

(6.) Khurana KK, Richards VI, Chopra PS, et al. The role of ultrasonography-guided fine-needle aspiration biopsy in the management of nonpalpable and palpable thyroid nodules. Thyroid 1998;8: 511-15.

(7.) Shaha AR. Controversies in the management oF thyroid nodule. Laryngoscope 2000;110(2 Pt 1): 183-93.

(8.) Carmeci C, Jeffrey RB, McDougall IR, et al. Ultrasound-guided fine-needle aspiration biopsy of thyroid masses. Thyroid 1998;8: 283-9.

(9.) Rosen IB, Azadian A, Walfish PG, et al. Ultrasound-guided fine-needle aspiration biopsy in the management of thyroid disease. Am J Surg 1993;166:346-9.

(10.) McIvor NP, Freeman JL, Salem S, et al. Ultrasonography and ultrasound-guided fine-needle aspiration biopsy of head and neck lesions: A surgical perspective. Laryngoscope 1994:104(6 Pt 1):669-74.

(11.) Sanchez RB, vanSonnenberg E, D'Agostino HB, et al. Ultrasound guided biopsy of nonpalpable and difficult to palpate thyroid masses. J Am Coll Surg 1994;178:33-7.

(12.) Sack MJ, Weber RS, Weinstein GS, et al. Image-guided fine-needle aspiration of the head and neck: Five years' experience. Arch Otolarygol Head Neck Surg 1998;124:1155-61.

(13.) Dwarakanathan AA, Staren ED, D'Amore MJ, et al. Importance of repeat fine-needle biopsy in the management of thyroid nodules. Am J Surg 1993;166:350-2.

(14.) Ezzat S, Sarti DA, Cain DR, Braunstein GD. Thyroid incidentalomas. Prevalence by palpation and ultrasonography. Arch Intern Med 1994;154:1838-40.

(15.) Mazzaferri EL. Management of a solitary thyroid nodule. N Engl J Med 1993;328:553-9.

(16.) Dawson B, Trapp RG. Basic and Clinical Biostatistics. 3rd ed. St. Louis: Lange Medical Books/McGraw Hill, 2001:152-60.

From the Division of Pediatric Otolaryngology, University of Iowa Hospital and Clinic, Iowa City (Dr. Goudy), and the Division of Surgical Oncology, Department of Surgery, University of Louisville School of Medicine, and the Center for Advanced Surgical Technologies, Norton Hospital, Louisville, Ky. (Dr. Flynn).

Reprint requests: Michael B. Flynn, MD, Division of Surgical Oncology, M-10, Department of Surgery, School of Medicine, University of Louisville, 315 E. Broadway, Louisville, KY 40292. Phone: [phone]; fax: [phone]; e-mail: [e-mail]