Deep-Learning-Based Screening and Ancillary Testing for Thyroid Cytopathology.

Thyroid cancer is the most common malignant endocrine tumor. The key test to assess preoperative risk of malignancy is cytologic evaluation of fine-needle aspiration biopsies (FNABs). The evaluation findings can often be indeterminate, leading to unnecessary surgery for benign post-surgical diagnoses. We have developed a deep-learning algorithm to analyze thyroid FNAB whole-slide images (WSIs). We show, on the largest reported data set of thyroid FNAB WSIs, clinical-grade performance in the screening of determinate cases and indications for its use as an ancillary test to disambiguate indeterminate cases. The algorithm screened and definitively classified 45.1% (130/288) of the WSIs as either benign or malignant with risk of malignancy rates of 2.7% and 94.7%, respectively. It reduced the number of indeterminate cases (N = 108) by reclassifying 21.3% (N = 23) as benign with a resultant risk of malignancy rate of 1.8%. Similar results were reproduced using a data set of consecutive FNABs collected during an entire calendar year, achieving clinically acceptable margins of error for thyroid FNAB classification.

Noninvasive follicular thyroid neoplasm with papillary-like nuclear features and the risk of malignancy in The Bethesda System for the Reporting of Thyroid Cytopathology.

INTRODUCTION: The Bethesda System for the Reporting of Thyroid Cytopathology (TBSRTC) is used to categorize and diagnose thyroid nodules by fine needle aspiration biopsy (FNAB). Each category in TBSRTC is associated with an estimated risk of malignancy (ROM). A subset of noninvasive encapsulated follicular variant of papillary thyroid carcinoma (niEFVPTC) was reclassified as a nonmalignant tumor: noninvasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP). We studied the impact of this reclassification on the reported ROM in TBSRTC. MATERIAL AND METHODS: We searched our institutional files for thyroid FNAB with surgical follow-up. ROM for each TBSRTC category was calculated. Subsequently, cases of niEFVPTC were reviewed and reclassified as NIFTP, if appropriate. ROM for each category was then recalculated after the reclassification. RESULTS: Twenty-six NIFTP were identified; the corresponding FNABs were distributed among all six TBSRTC categories. The majority of NIFTP FNAB were in the AUS/FLUS and suspicious for malignancy (SUSP) categories, 12 (46.2%) and 8 (30.8%), respectively. While the ROM changed for all diagnostic categories, the greatest change in ROM after reclassification was seen in these two categories. Absolute ROM for AUS/FLUS decreased from 25.0% to 21.0% and SUSP, from 71.7% to 58.3%, changes that were statistically significant. CONCLUSIONS: The reclassification of niEFVPTC to NIFTP has significantly impacted the ROM in the TBSRTC at our institution. While there was a decrease in ROM for all categories, the greatest reduction to ROM was in the categories of AUS/FLUS and FN. These changes to the ROM should help guide surgical approach moving forward.

Application of a machine learning algorithm to predict malignancy in thyroid cytopathology.

BACKGROUND: The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC) comprises 6 categories used for the diagnosis of thyroid fine-needle aspiration biopsy (FNAB). Each category has an associated risk of malignancy, which is important in the management of a thyroid nodule. More accurate predictions of malignancy may help to reduce unnecessary surgery. A machine learning algorithm (MLA) was developed to evaluate thyroid FNAB via whole slide images (WSIs) to predict malignancy. METHODS: Files were searched for all thyroidectomy specimens with preceding FNAB over 8 years. All cytologic and surgical pathology diagnoses were recorded and correlated for each nodule. One representative slide from each case was scanned to create a WSI. An MLA was designed to identify follicular cells and predict the malignancy of the final pathology. The test set comprised cases blindly reviewed by a cytopathologist who assigned a TBSRTC category. The area under the receiver operating characteristic curve was used to assess the MLA performance. RESULTS: Nine hundred eight FNABs met the criteria. The MLA predicted malignancy with a sensitivity and specificity of 92.0% and 90.5%, respectively. The areas under the curve for the prediction of malignancy by the cytopathologist and the MLA were 0.931 and 0.932, respectively. CONCLUSIONS: The performance of the MLA in predicting thyroid malignancy from FNAB WSIs is comparable to the performance of an expert cytopathologist. When the MLA and electronic medical record diagnoses are combined, the performance is superior to the performance of either alone. An MLA may be used as an adjunct to FNAB to assist in refining the indeterminate categories.

Treatment Outcomes in Anaplastic Thyroid Cancer.

BACKGROUND: Anaplastic thyroid cancer (ATC) is rare, accounting for 1-2% of thyroid malignancies. Median survival is only 3-10 months, and the optimal therapeutic approach has not been established. This study aimed to evaluate outcomes in ATC based on treatment modality. METHODS: Retrospective review was performed for patients treated at a single institution between 1990 and 2015. Demographic and clinical covariates were extracted from the medical record. Overall survival (OS) was modeled using Kaplan Meier curves for different treatment modalities. Univariate and multivariate analyses were conducted to assess the relationships between treatment and disease characteristics and OS. RESULTS: 28 patients with ATC were identified (n = 16 female, n = 12 male; n = 22 Caucasian, n = 6 African-American; median age 70.9). Majority presented as Stage IVB (71.4%). Most patients received multimodality therapy. 19 patients underwent local surgical resection. 21 patients received locoregional external beam radiotherapy (EBRT) with a median cumulative dose of 3,000 cGy and median number of fractions of 16. 14 patients received systemic therapy (n = 11 concurrent with EBRT), most commonly doxorubicin (n = 9). 16 patients were never disease free, 11 patients had disease recurrence, and 1 patient had no evidence of disease progression. Median OS was 4 months with 1-year survival of 17.9%. Regression analysis showed that EBRT (HR: 0.174; 95% CI: 0.050-0.613; p=0.007) and surgical resection (HR: 0.198; 95% CI: 0.065-0.598; p=0.004) were associated with improved OS. Administration of chemotherapy was not associated with OS. CONCLUSIONS: Anaplastic thyroid cancer patients receiving EBRT to the thyroid area/neck and/or surgical resection had better OS than patients without these therapies, though selection bias likely contributed to improved outcomes since patients who can undergo these therapies tend to have better performance status. Prognosis remains poor overall, and new therapeutic approaches are needed to improve outcomes.