Quality Improvement by Standardization of Procurement and Processing of Thyroid Fine-Needle Aspirates in the Absence of On-site Cytological Evaluation

Introduction

C linical history, physical examination, laboratory studies, and various imaging modalities are employed when triaging thyroid nodules (TNs) that require fine-needle aspiration (FNA). It is of fundamental importance to obtain adequate and interpretable/satisfactory samples for the cytopathologist to render a meaningful interpretation. This includes procuring adequate, representative material and producing good quality smears for cytopathological interpretation. Further, a meaningful diagnosis should employ standardized diagnostic terminology that facilitates clinical efforts to formulate therapeutic plans (1). The rate of nondiagnostic/unsatisfactory TN FNAs has been reported at 5–43% and TN FNA is inconclusive for neoplasia in 5–29% of the patients (2 –9). Mayall et al. showed a decrease in the overall inadequacy rate from 29% to 9% in FNAs from multiple sites when the biopsy and sample preparation processes were performed by a limited number of trained individuals (10). At our institution we experienced a high rate of unacceptable TN FNAs, which we felt was likely because the aspiration and preparative procedures were performed by a wide variety of physicians and nurses and/or technical staff.

We implemented a major change in these methods by limiting the performance of all TN FNAs to one physician and selected support staff specifically trained in TN FNA collection, and by standardizing preparation of the samples for cytopathological review. We hypothesized that limiting the number of individuals involved and having them prepare samples according to a well-defined standardized method would increase the proportion of samples adequate for cytopathological review and diagnosis.

Materials and Methods

This is a retrospective, before-and-after cohort study that examined the improvement in TN FNA samples satisfactory for cytopathological examination and diagnosis following introduction of standardized methods for TN FNA sample collection and preparation. In the prestandardization period, TN FNAs were performed by multiple physicians (surgeons, endocrinologists, and radiologists) with variable use of ultrasound guidance and variably sized needles (22–27 gauge). Smears (some air-dried and/or fixed, 2–25 smears per FNA) were made by numerous individuals with variable training and experience in these methods. Standardization of collection was instituted on June 1, 2006 following the collaborative efforts of pathology, endocrinology, and appropriate support staff. Poststandardization TN FNAs were performed by one endocrinologist under ultrasound guidance, using a 25-gauge needle. A total of four passes were made. The initial two passes resulted in four standard smears (one air-dried and one fixed, per pass) prepared by a single designated and trained nurse, with needle rinses and additional pass material submitted in CYTORICH^® Red Preservative Fluid (BD Diagnostics TriPath, Burlington, NC) for BD SurePath™ (BD Diagnostics TriPath) liquid-based preparation and/or cell block and hematoxylin and eosin studies. Standard Papanicolaou and Diff–Quick staining techniques were used in all cases.

The study was submitted to the Institutional Review Board at Dartmouth College/Dartmouth Hitchcock Medical Center and was determined to be exempt from review. All TN FNAs performed between October 1, 2004 and August 29, 2007 were identified by a computerized archival search of our files. Cases were classified into prestandardization (before June 1, 2006) or poststandardization (after June 1, 2006) periods. Pathology reports were reviewed for specimen quality and cytopathological diagnosis. Patient follow-up (f/u) data were also collected.

Specimen quality was classified as satisfactory, unsatisfactory, or less than optimal (LTO). A satisfactory specimen was defined by the presence of at least five to six cell groups each containing at least 10 preserved cells on at least two smears and devoid of obscuring factors (9). LTO was defined as smears where a cytological interpretation could be rendered, but where the material was compromised in some way, for example, abundant/obscuring blood, scant follicular cells, but abundant colloid and/or macrophages, some air-drying of the fixed smears. Unsatisfactory was defined as smears where a cytological interpretation could not be rendered, for example, on smears with nil to scant material and/or obscuring blood.

Cytopathological diagnoses were evaluated as benign, atypical, suspicious for malignancy, or positive for malignancy using a pattern approach (11). The “benign” category included thyroiditis, multinodular goiters, and adenomatoid nodules including smears with predominantly macrofollicular patterns or mixed macro- and microfollicular patterns with abundant background colloid and cases with adequate cellularity and features compatible with cyst contents. The “atypical” category included smears with low cellularity and predominantly three-dimensional groups and/or microfollicular pattern and/or cytological atypia and those cases with low cellularity and/or obscuring blood where papillary carcinoma could not be excluded. The atypical designation communicated that f/u and/or reaspiration were warranted. The “suspicious for malignancy” category included smears with predominance of microfollicular pattern suggestive of follicular neoplasm, predominance of Hurthle cells suggestive of Hurthle cell neoplasm, and smears with some nuclear features suggestive but not diagnostic of papillary carcinoma. The “positive for malignancy” category included cases that exhibited definitive criteria of malignancy, including cases of papillary carcinoma, medullary carcinoma, metastasis, and lymphoma.

Patient f/u parameters were collected as secondary outcome data and classified as medical, surgical, or repeat FNA. Data gathered in all cases that resulted in surgery included the following: the participating surgeon, the procedure (i.e., hemithyroidectomy vs. complete thyroidectomy), and the histopathological findings.

Results

One hundred patients were included in the prestandardization group, with a total of 121 TN FNAs. Ninety-one patients were included in the poststandardization group, with a total of 143 TN FNAs (Table 1). No patients identified by the computer-assisted search were excluded from the study. There were nine repeat FNAs in the prestandardization group and 13 repeat FNAs in the poststandardization group. Specimen quality, as judged by the reviewing pathologist, was considered an ordered outcome of satisfactory, LTO, or unsatisfactory. Standardization of TN FNA and sample preparation resulted in an overall improvement in the quality of the samples (odds ratio = 3.82, 95% CI 2.02–7.24, p < 0.0001) (Table 2). This is seen with the increased proportion of satisfactory samples from the 67% observed prestandardization period to 89% following standardization.

Secondary outcomes in the prestandardization group revealed that all of the 13 (11%) unsatisfactory cases were due to obscuring blood. Of the 27 (22%) LTO cases, 93% were due to hypocellularity and cystic degeneration. Other cited factors for LTO were obscuring blood and air-drying artifact. Of the 81 (67%) satisfactory cases, 45 (56%) were benign, 21 (26%) atypical, 14 (17%) suspicious for malignancy, and 1 (1%) was positive for malignancy. The surgical pathological findings based on surgical f/u of 18/21 atypical cases (2 hemithyroidectomies, 16 complete thyroidectomies), 14/14 suspicious for malignancy cases (1 hemithyroidectomy, 13 complete thyroidectomies), and 1/1 positive for malignancy cases (complete thyroidectomy) are shown in Table 3. Five cases with benign diagnoses underwent complete thyroidectomy revealing five multinodular goiters, one of which contained an incidental papillary carcinoma. The remaining benign cases were followed medically.

Secondary outcomes in the poststandardization group revealed that of the six (4%) unsatisfactory cases, half were due to obscuring blood and half were due to hypocellularity. All of the 10 (7%) LTO cases were due to hypocellularity, with comments in 8 (80%) of these cases that the specimen likely represented cyst contents. Of the 127 (89%) satisfactory cases, 93 (73%) were benign, 25 (20%) were atypical, 7 (5%) were suspicious for malignancy, and 2 cases (2%) were positive for malignancy. There was an increase in the percentage of benign cases from satisfactory samples poststandardization (56–73%; odds ratio = 2.17, 95% CI 1.16–4.11, p = 0.0135).

The surgical pathological findings based on surgical f/u of 24/25 atypical cases (4 hemithyroidectomies, 20 complete thyroidectomies), 7/7 suspicious for malignancy cases (complete thyroidectomy), and 2/2 positive for malignancy cases (complete thyroidectomy) are shown in Table 3. Of the benign cases, 12 underwent complete thyroidectomy. Ten of the 12 benign nodules underwent surgery because of additional atypical lesions within those glands, and the other two cases resulted in surgery because the patient's had symptomatic multinodular goiters. The remaining benign cases were followed medically.

Discussion

TNs are common and are present in approximately 50% of adults with clinically normal thyroid glands (12 –14). Given that the majority of TNs are benign, TN FNA is used to help stratify patients according to the risk of malignancy and facilitate the selection of higher risk patients for surgery. It is an effective management tool in this setting, with a reported diagnostic accuracy rate of 90–100% (15,16). Specimens must be adequately representative of the lesion, as persistently inadequate specimens do not permit adequate cytopathological review and may precipitate unnecessary surgery (17,18). TN FNA specimens should yield a low false-negative rate so that malignant lesions are not missed. Reported false-negative rates range from 1% to 11% and mostly are due to unsatisfactory specimens (19).

Specimen adequacy can be viewed in terms of quantity and quality. The quantity aspect of adequacy is controversial and nonstandardized, whereas the quality aspect of adequacy should not be regarded as controversial (1,20). Optimal quality TN FNA specimens should reflect technical adequacy and exhibit good preservation and preparation with adequate, interpretable staining. On-site cytological assessment of the TN FNA samples by a cytotechnologist and/or a pathologist is not always feasible and is not performed in our institution. Our hypothesis was that our standardization measures would result in better quality specimens and ultimately improve interpretability even in the absence of on-site cytological assessment of the specimen. We found that standardization markedly improved the proportion of satisfactory TN FNA specimens, reducing both the number of unsatisfactory and LTO cytopathological interpretations. In the prestandardization period, the chance of receiving and evaluating an unsatisfactory or LTO specimen was approximately four times greater than it was following standardization. Improved specimen quality was reflected in better specimen interpretability and more consistent rendering of meaningful diagnoses, translating into better patient care by decreasing unnecessary surgery, repeat procedures, and potential associated delays in surgical management. The increase in the percentage of benign cases from satisfactory samples poststandardization (56–73%; odds ratio = 2.17, 95% CI 1.16–4.11, p = 0.0135) is another indicator of the improvement in FNA methods.

Many types of physicians with various levels of training perform TN FNAs. Standardization of the process in our institution required persistence. Currently, the ideal physician to perform thyroid FNA in an institutional or office practice should be one who is experienced and has repeatedly demonstrated appropriate judgment in nodule selection, technical excellence, and proficiency in obtaining aspirate material and preparing slides (21). The best results are achieved when either a single physician or a team of physicians trained and experienced in TN FNA perform the entire procedure (22), as we demonstrated in this study. However, as this may not be a practical option, we feel that it is essential to establish standardization of techniques, especially the size of the needle, number of passes, number of smears, and fixation for that institution/group with pathologist preference in mind. This requires persistence, a willingness to educate, and communication on the part of the cytopathologist.

In conclusion, standardization of the TN FNA process has resulted in a significant improvement in specimen quality and interpretability in our institution in the absence of on-site cytological evaluation, subjectively making TN FNA interpretation a favorable experience.