Key Points

• Assessed molecular testing techniques allowed 49% of patients with indeterminate thyroid nodules to avoid diagnostic surgery.
• The 2 commercially available tests compared in this study (an RNA test and a DNA‐RNA test) had similar sensitivity and specificity.

A new study compared the performance characteristics of an RNA test and a DNA‐RNA test for predicting the malignancy risks of the approximately 20% of thyroid nodules that display indeterminate fine‐needle aspiration (FNA) cytology results. The researchers found that each test allowed nearly half of patients with indeterminate nodules to opt for observation to avoid diagnostic surgery and the possibility of lifelong thyroid hormone supplementation following surgery and to alleviate stress. The study was published in JAMA Oncology (2021;7:70‐77. doi:10.1001/jamaoncol.2020.5935).

Study author Masha J. Livhits, MD, an endocrine surgeon and assistant professor of surgery at the David Geffen School of Medicine at the University of California Los Angeles, says that the study breaks new ground because it is the first and only direct comparison of the 2 leading commercially available molecular test technologies that clinicians use to help to determine whether cytologically indeterminate thyroid nodules should be observed or undergo hemithyroidectomy. Livhits explained, “Prior studies are limited by evaluating only one test (not both), and it is difficult to compare the performance of tests across studies as that performance is dependent on the prevalence of malignancy in a given population.”

Study Details

Patients in this randomized clinical trial came from the UCLA Health system and underwent thyroid FNA biopsies between August 1, 2017, and November 30, 2019. Only patients with at least 1 cytologically indeterminate nodule according to categories III and IV of the Bethesda System for Reporting Thyroid Cytopathology were included.

The FNA biopsies were performed by radiologists, endocrinologists, and endocrine surgeons at 9 clinical sites. These specimens were analyzed by experienced head and neck cytopathologists and also sent for molecular testing if their cytopathology was indeterminate. Of 427 indeterminate nodules from 397 patients, 55 (from 51 patients) were excluded according to criteria such as a concurrent malignant cytology finding from another thyroid nodule from the same patient or physician preference for one of the molecular tests rather than inclusion in a randomized study. The 372 included nodules (from 346 patients) were block‐randomized by month to undergo testing with either the RNA test or the DNA‐RNA test.

Molecular test results (classified as benign or suspicious for malignancy) were compared with the corresponding surgical specimen diagnoses from expert thyroid pathologists, although nodules with benign molecular test results were assumed to be truly benign even if they were not resected if they had stable follow‐up surveillance ultrasounds. Nodules with noninvasive follicular thyroid neoplasm with papillary‐like features were grouped with cancer “because it represents a premalignant entity that should be managed surgically,” wrote the researchers. The study’s primary outcomes were the sensitivity, specificity, positive predictive value, and negative predictive value for each of the molecular tests.

Study Results

Of the 372 cytologically indeterminate nodules that were included for analysis, 54% (n = 201) were randomized to undergo the RNA test, and 46.0% (n = 171) were randomized for the DNA‐RNA test. The patients’ median age was 55 years, and 76.9% were female (n = 266).

The RNA test demonstrated benign results in 107 nodules (53.2%) and suspicious results in 73 nodules (36.3%). There were 19 samples (9.5%) judged to be inadequate for analysis and 2 samples (1.0%) that were suspicious for parathyroid (rather than thyroid) tissue. Of the 107 benign nodules tested with the RNA test, 12 (11.2%) were surgically resected. Of the 73 nodules with suspicious RNA test results, 58 (79.5%) were resected.

The DNA‐RNA test demonstrated negative results in 103 nodules (60.2%) and positive results in 60 nodules (35.1%). There were 7 samples (4.1%) that were insufficient for molecular analysis and 1 sample (0.6%) that was suspicious for being a sample of parathyroid tissue. Of the nodules with negative results from the DNA‐RNA test, 11 (10.7%) were surgically resected. Of the nodules with positive results from the DNA‐RNA test, 49 (81.7%) were resected.

In analyses including comparisons with either surgical or clinical surveillance results, the RNA test and the DNA‐RNA test had similar results (with P values ranging from .32 to >.99) for sensitivity (100.0% and 96.9%, respectively), specificity (79.6% and 84.8%, respectively), positive predictive value (53.5% and 63.3%, respectively), and negative predictive value (100.0% and 99.0%, respectively). All 4 performance characteristics were also statistically similar in comparisons with surgical results only.

Use of either test allowed 49% of patients with indeterminate thyroid nodules to avoid diagnostic surgery.

Study Analysis

“While the 2 most commonly used molecular tests for thyroid nodules are the RNA‐based Afirma test [Veracyte, Inc., San Francisco, California] and the DNA‐ and RNA‐based ThyroSeq v3 test [University of Pittsburgh Medical Center and Sonic Healthcare USA], it is not clear which test is better,” says William C. Faquin, MD, PhD, professor of pathology at Harvard Medical School and director of head and neck pathology at Massachusetts Eye and Ear Infirmary in the Divisions of Head and Neck Pathology and Cytopathology at Massachusetts General Hospital in Boston. “The importance of this study is that it provides a clear‐cut comparison using a parallel randomized clinical trial to assess the performance of both of these molecular testing options. Both the Afirma and ThyroSeq v3 tests performed equally well in this role. The current study in JAMA Oncology is very helpful in synthesizing a lot of the data, which can be confusing, and providing a clear message about the strong performances of both tests.”

Dr. Livhits says that the key takeaway message is that molecular analysis with either test can help patients to avoid unnecessary diagnostic surgery. “Our results indicate that the two tests have similar diagnostic performance, with very high sensitivity and moderate specificity,” she says. “Therefore, the choice of which test to use may be based on other factors, such as the cost of the test and the turnaround time. This is important given the increasing detection of thyroid nodules and the need to avoid overtreatment. ThyroSeq v3 provides data on specific mutations, which were reported in our study, that may impact the prognosis of nodules ultimately found to be malignant and may guide the extent of treatment or provide therapeutic targets in the rare advanced, metastatic cases.”

For clinicians seeking more guidance on this topic, Dr. Faquin recommends the American Thyroid Association Guidelines (thyroid.org/professionals/ata‐professional‐guidelines/). “They address in detail the application of molecular testing in the cytologic evaluation of thyroid nodules. The topic is also covered very well in several review articles in the cytology and endocrinology literature.”

Photo credit: Cancer Cytopathol. 2011;120:73‐86. doi:10.1002/cncy.20178

中文翻译：

---

两项领先的甲状腺分子检测可能会减少对诊断手术的需求

关键点

• 经评估的分子检测技术使 49% 的甲状腺结节不确定的患者避免了诊断性手术。
• 本研究中比较的两种市售测试（RNA 测试和 DNA-RNA 测试）具有相似的敏感性和特异性。

一项新研究比较了 RNA 测试和 DNA-RNA 测试的性能特征，以预测大约 20% 的甲状腺结节的恶性风险，这些甲状腺结节显示不确定的细针穿刺 (FNA) 细胞学结果。研究人员发现，每次测试都允许近一半患有不确定结节的患者选择观察，以避免诊断性手术和手术后终身补充甲状腺激素的可能性，并减轻压力。该研究发表在JAMA Oncology (2021;7:70-77.doi:10.1001/jamaoncol.2020.5935)。

研究作者、加州大学洛杉矶分校大卫格芬医学院内分泌外科医生兼外科助理教授、医学博士 Masha J. Livhits 表示，这项研究开辟了新天地，因为它是第一个也是唯一一个直接比较2 种领先的商用分子检测技术，临床医生使用这些技术来帮助确定是否应观察细胞学上不确定的甲状腺结节或进行半甲状腺切除术。Livhits 解释说，“先前的研究受到仅评估一项测试（而不是两者）的限制，并且很难比较不同研究的测试性能，因为该性能取决于特定人群中恶性肿瘤的患病率。”

学习详情

这项随机临床试验的患者来自加州大学洛杉矶分校卫生系统，并在 2017 年 8 月 1 日至 2019 年 11 月 30 日期间接受了甲状腺 FNA 活检。根据 Bethesda 报告系统的 III 和 IV 类，只有具有至少 1 个细胞学不确定结节的患者包括甲状腺细胞病理学。

FNA 活检由放射科医师、内分泌科医师和内分泌外科医师在 9 个临床中心进行。这些样本由经验丰富的头颈部细胞病理学家进行分析，如果它们的细胞病理学不确定，也会送去进行分子检测。在来自 397 名患者的 427 个不确定结节中，根据标准排除了 55 个（来自 51 名患者），例如同一患者的另一个甲状腺结节同时发现恶性细胞学或医生偏好其中一种分子检测而不是纳入随机研究. 纳入的 372 个结节（来自 346 名患者）按月分组随机进行 RNA 测试或 DNA-RNA 测试。

将分子检测结果（分类为良性或疑似恶性肿瘤）与专家甲状腺病理学家的相应手术标本诊断进行比较，尽管分子检测结果为良性的结节被认为是真正良性的，即使它们没有被切除，如果它们有稳定的随访。起来监视超声波。研究人员写道，具有乳头状特征的非侵袭性甲状腺滤泡性肿瘤结节被归类为癌症，“因为它代表了一种应该通过手术治疗的癌前实体”。该研究的主要结果是每个分子检测的敏感性、特异性、阳性预测值和阴性预测值。

研究结果

在纳入分析的 372 个细胞学不确定结节中，54% (n = 201) 随机接受 RNA 测试，46.0% (n = 171) 随机接受 DNA-RNA 测试。患者的中位年龄为 55 岁，76.9% 为女性（n = 266）。

RNA 检测显示 107 个结节 (53.2%) 为良性结果，73 个结节 (36.3%) 为可疑结果。有 19 个样本 (9.5%) 被判定为不适合分析，2 个样本 (1.0%) 怀疑有甲状旁腺（而不是甲状腺）组织。在用 RNA 测试测试的 107 个良性结节中，有 12 个（11.2%）被手术切除。RNA 检测结果可疑的 73 个结节中，切除了 58 个（79.5%）。

DNA-RNA 检测显示 103 个结节 (60.2%) 为阴性结果，60 个结节 (35.1%) 为阳性结果。有7个样本（4.1%）不足以进行分子分析，1个样本（0.6%）怀疑是甲状旁腺组织样本。在 DNA-RNA 检测结果为阴性的结节中，有 11 个（10.7%）被手术切除。在 DNA-RNA 检测结果为阳性的结节中，有 49 个（81.7%）被切除。

在包括与手术或临床监测结果比较在内的分析中，RNA 测试和 DNA-RNA 测试在敏感性（分别为 100.0% 和 96.9%）、特异性方面具有相似的结果（ P值范围从 0.32 到 >.99） （分别为 79.6% 和 84.8%）、阳性预测值（分别为 53.5% 和 63.3%）和阴性预测值（分别为 100.0% 和 99.0%）。仅与手术结果相比，所有 4 个性能特征在统计学上也相似。

使用任何一种测试都可以让 49% 的甲状腺结节不确定的患者避免进行诊断性手术。

研究分析

“虽然 2 种最常用的甲状腺结节分子检测是基于 RNA 的 Afirma 检测 [Veracyte, Inc., San Francisco, California] 和基于 DNA 和 RNA 的 ThyroSeq v3 检测 [匹兹堡大学医学中心和 Sonic Healthcare美国]，目前尚不清楚哪种测试更好，”哈佛医学院病理学教授兼马萨诸塞州眼耳医院头颈科头颈病理学主任 William C. Faquin 医学博士说波士顿马萨诸塞州总医院的病理学和细胞病理学。“这项研究的重要性在于，它使用平行随机临床试验提供了明确的比较，以评估这两种分子测试选项的性能。Afirma 和 ThyroSeq v3 测试在这个角色中表现同样出色。目前的研究在JAMA Oncology在合成大量可能令人困惑的数据方面非常有帮助，并提供了关于两种测试的强劲表现的明确信息。”

Livhits 博士说，关键的外卖信息是，任何一种测试的分子分析都可以帮助患者避免不必要的诊断手术。“我们的结果表明，这两种测试具有相似的诊断性能，具有非常高的灵敏度和中等的特异性，”她说。“因此，选择使用哪种测试可能基于其他因素，例如测试成本和周转时间。鉴于甲状腺结节的检出率越来越高，并且需要避免过度治疗，这一点很重要。ThyroSeq v3 提供了我们研究中报告的特定突变的数据，这些突变可能会影响最终发现为恶性的结节的预后，并可能指导治疗范围或为罕见的晚期转移性病例提供治疗靶点。”

对于在该主题上寻求更多指导的临床医生，Faquin 博士推荐美国甲状腺协会指南 (甲状腺.org/professionals/ata-professional-guidelines/)。“他们详细介绍了分子检测在甲状腺结节细胞学评估中的应用。细胞学和内分泌学文献中的几篇评论文章也很好地涵盖了该主题。”

图片来源：癌症细胞病理学。2011；120:73-86。doi:10.1002/cncy.20178