Precision Medicine Comes to Thyroidology

Abstract

Thyroidology has long been in the vanguard of discovery in endocrine physiology, pathophysiology, and clinical medicine. The structure of T₄ was described before those of cortisol and testosterone (1), as was the trophic role of the anterior pituitary in regulating thyroid gland function. T₃ receptors were discovered almost simultaneously with those for steroid hormones (2). Thyroid hormone replacement therapy (3) was preceded in clinical endocrinology only by that for hypogonadism. And surgical and medical treatments for thyroid hyperfunction preceded those for disorders of the other endocrine glands. Diagnostic and therapeutic use of radioiodine gave birth to the entire field of nuclear medicine. As a result of these and subsequent advances, there are now highly accurate diagnostic tests and reasonably effective therapeutic options for disorders of thyroid function and the vast majority of thyroid tumors.

Despite these achievements, the clinical solutions offered in contemporary thyroid practice have their shortcomings: uncertainties about whether subtle changes in thyroid function and structure actually represent diseases warranting treatment, the dissatisfaction that a minority of hypothyroid patients have with their response to hormone replacement, the hope of treating hyperthyroidism permanently but without destroying or removing the gland, the overtreatment of many patients with thyroid nodules and even thyroid cancers, and the unmet need to improve the management of patients with aggressive thyroid malignancies. There are also unrealized opportunities to address certain nonthyroidal disorders with organ-specific thyroid hormone agonists.

In response to the emphasis on precision medicine (4) in the decade ahead, we can anticipate that advances in thyroid pathophysiology will help us refine the assessment of individual patients to determine whether their thyroid function test or imaging findings really represent conditions with clinical significance for their health. We can also expect development of management strategies better targeted to each patient's needs and likelihood of having a salutary response. The following sections describe examples of those individual needs and the precision solutions that we can reasonably hope will emerge over the next decade as a result of discoveries in basic and clinical thyroidology.

Hypothyroidism

Despite great success in accurately diagnosing and effectively treating primary hypothyroidism—largely the result of precise TSH immunoassays—there remain uncertainties about which patients with mild hypothyroidism should be treated (5) and why a minority of those who are treated with levothyroxine remain dissatisfied with their health (6). Decisions about whether to screen for and treat patients with an elevated TSH and normal free T₄ are already better informed than they were a decade ago. We now appreciate that there is a normal age-related increase in the distribution of serum TSH values among the elderly with no goiter, thyroid autoantibodies, or risk factors for thyroid dysfunction (7). We have a body of epidemiological (8) and nonrandomized trial data (9) indicating that any cardiovascular benefit of thyroid hormone therapy for subclinical hypothyroidism appears limited to middle-aged, but not elderly, adults. We also know that levothyroxine treatment of mild hypothyroidism—whether spontaneously occurring (10) or artificially created (11)—fails to improve most measures of neurocognitive function and mood. And we appreciate that isolated TSH elevation often reverts to normal, even among those with circulating thyroid autoantibodies (12). Finally, we have more data demonstrating the high incidence and significant consequences of iatrogenic thyrotoxicosis in levothyroxine-treated patients (13). By the end of the next decade, we will finally have what should be definitive evidence from a large prospective randomized clinical trial of levothyroxine treatment for subclinical hypothyroidism (14). This should more clearly inform clinical decision-making about management for this huge population. And the results might be surprisingly negative, as they were earlier in a large well-designed study of treatment for subclinical hypothyroidism in pregnant women (15).

It is less clear that the next decade will provide definitive answers to help those patients with treated hypothyroidism whose expectations have not been fulfilled. The notion that augmenting levothyroxine with T₃ is generally superior has been rather soundly refuted by the vast majority of randomized blinded trials (16), unless combination therapy induces iatrogenic thyrotoxicosis. In the decade ahead, we can look forward to more thorough testing of the hypothesis that variations in deiodinase genes, protein structure, and activity might explain the dissatisfaction of some levothyroxine-treated hypothyroid patients, such as those reported in one study to have favorable responses to T₄ plus T₃ therapy (17). If this observation was substantiated, it would represent a splendid application of precision medicine to the management of hypothyroidism. It is also possible that future research will demonstrate that other iodothyronines or their metabolites, such as the thyronamines (18), might have important physiological roles and therapeutic applications.

Future clinical practice will no doubt include the use of devices offering patients with treated hypothyroidism greater autonomy in monitoring their status, such as adaptation of biosensors to provide warnings that altered heart rate, body weight, temperature, and activity and sleep patterns might indicate suboptimal thyroid hormone therapy and the need for home TSH testing on a finger stick blood specimen.

Thyrotoxicosis

In contemporary practice, highly accurate laboratory tests recognize even the mildest degrees of thyrotoxicosis, and there are three effective therapeutic modalities to treat its most common causes. In the next decade, it seems likely that biomedical science will provide additional options. First, preclinical studies have already described small molecule inhibitors of TSH receptor binding and/or activation (19), which could be effective in inhibiting its stimulation by antibodies in Graves' disease, gonadotropins in chorionic tumors, and familial gestational thyrotoxicosis, and TSH itself in centrally mediated hyperthyroidism. Second, the success of immunomodulatory biological agents in treating autoimmune connective tissue disorders foreshadows future investigation and perhaps clinical applications of similar agents to induce remission in patients with autoimmune hyperthyroidism. However, for both of these novel pharmacological approaches, there will be significant barriers to widespread clinical use of agents that will be much more expensive, associated with their own adverse effects, and probably no more likely than thionamides to be associated with permanent remissions once their use is discontinued. Consequently, even if these innovative approaches are effective, the key to justifying their use will be precision in identifying those hyperthyroid Graves' patients with a prognostic signature (eg, HLA or CTLA-4 haplotypes, impaired regulatory T cells, or inappropriate HLA-DR expression) predicting greater risk of failure with conventional antithyroid drugs or, in the case of immunomodulatory drugs, a greater likelihood of inducing a sustained remission. Graves' ophthalmopathy represents, of course, an even more compelling unmet medical need, but the lack of a more complete understanding of its pathogenesis along with the challenge of enrolling a sufficient number of appropriate patients in clinical trials will make this a particularly challenging disorder to address.

There is also the opportunity to improve the use of radioiodine to treat hyperthyroid patients with Graves' disease, toxic adenoma, and toxic multinodular goiter. 131-I doses have traditionally been determined empirically or after rudimentary dosimetry based on a clinical estimate of gland or nodule volume and a thyroidal radioiodine uptake measurement at a single time point. Precision 131-I therapy could apply more sophisticated dosimetry with single photon emission computed tomography (SPECT)-CT fusion imaging or even 124-I positron emission tomography (PET)-CT to estimate functional gland volume (20) and with area-under-the-curve radioisotope retention data acquired with a monitoring collar capable of remote communication after administering a tracer dose (21).

Thyroid Nodules and Cancer

Increased recognition of thyroid nodules and cancers in recent years has generated debate about the degree to which this represents overdiagnosis due to the combination of highly sensitive imaging and imperfect preoperative testing to differentiate benign from malignant nodules (22). The consequence has been a parallel increase in performance of thyroid surgeries (23), most of which prove, in retrospect, to have been for benign nodules. During the past decade, implementation of standardized criteria for thyroid nodule cytopathology and introduction of molecular diagnostic testing on cytological samples have gone a long way to reduce unnecessary thyroid surgeries with their complications and cost (24).

At the same time, however, we have learned that there is a disquieting discordance between even expert thyroid pathologists about which surgical specimens are, in fact, malignant (25). Furthermore, a new perspective about encapsulated follicular variant of papillary cancers (26)—now proposed to be benign—has raised additional uncertainty about the accuracy of conventional histopathology in predicting thyroid tumor behaviors. Finally, early results of a controlled trial comparing observation with surgery for papillary microcarcinomas indicate that most of these tumors can be safely monitored without resection (27).

The decade of precision medicine ahead will move clinicians beyond reliance on microscopic and molecular findings for diagnosis to applying these observations to prognostication—tailoring management, whether surgery or observation, to achieve the best outcome for the individual thyroid cancer patient. We now recognize that certain oncogenic markers present in a minority of thyroid cancers, such as BRAF and TERT mutations, can be powerful predictors of aggressive tumor behaviors, such as extrathyroidal spread and nodal metastases, leading to increased tumor recurrence and reduced survival (28). Just as the presence of certain gene expression patterns now enables us to recognize benign thyroid nodules (29), the absence of other molecular indicators of aggressiveness should enable clinicians to observe many thyroid nodule patients—even those whose histopathology might previously have been categorized as malignant. The integrated genomic characterization of papillary thyroid cancer (30) demonstrates how combining analyses of genomic variants, gene expression, and methylation patterns defines distinct driver groups with characteristic signaling and differentiation profiles.

Precision medicine will also refine our approaches to radiotherapy and chemotherapy for those thyroid cancer patients whose tumors require more intensive treatment. Our greater understanding of sodium-iodide transporter function, how its expression is regulated, and how it is inserted into the cell membrane may lead to interventions that restore iodine avidity to the more aggressive thyroid cancers, which often lose it (see below). Previous approaches to radioiodine dosimetry have been either empiric or based on dose-limiting whole body or blood exposure. 124-I PET-CT and 131-I PET-CT have the potential to accurately predict lesional dosimetry (31), offering the promise of more effective radioiodine therapy for metastases, avoidance of futile retreatments, and minimization of adverse effects.

Precision mutational targeting of thyroid cancer chemotherapy is already a work in progress. For example, the BRAF mutated genotype predicts better responsiveness to combined BRAF and MEK inhibition, such as with dabrafenib and trametinib. A particularly exciting development has been the demonstration that selumetinib can redifferentiate a subset of metastatic thyroid cancers with poor or no radioiodine avidity—rendering them treatable with 131-I (32). This response has been shown to occur preferentially in patients whose tumors harbor NRAS and BRAF mutations. Hopefully, advances like these will also reveal new chemotherapeutic targets for follicular thyroid cancer, such as the PAX8-PPARγ fusion oncoprotein (33), for which the familiar PPARγ agonist pioglitazone is being evaluated in a clinical trial. For the dreaded anaplastic thyroid cancer, an integrated genomic characterization should identify new targets, such as striatin-Alk fusions (34), which can be inhibited with crizotinib. Finally, the novel immunomodulatory drugs producing dramatic responses in some patients with other malignancies will certainly be further investigated and may prove useful in some aggressive thyroid cancers. All of these developments herald the dawn of an era when an individual thyroid cancer patient's tumor can be genetically characterized to select the most promising chemotherapeutic option.

Thyroid Hormone Treatments for Nonthyroidal Disorders

Precision thyroid agonists will be further studied as a potential treatment for other medical conditions, eg, targeting the liver for hypercholesterolemia (35) and nonalcoholic fatty liver disease, brown fat for obesity (36), and mitochondria for adrenoleukodystrophy (37). The challenge, of course, will remain avoiding off-target thyrotoxic effects, particularly affecting the skeleton, cartilage, and cardiac-conducting system. The design of drugs with preferential binding to the TRβ1 isoform of the thyroid hormone receptor, which predominates in liver, and lesser binding to the TRα1 isoform, which is expressed in the cardiac-conducting system, represents one approach. However, as knowledge grows about the tissue-specific distribution of thyroid hormone transporters and new thyromimetic compounds are developed with moieties facilitating selective organ uptake, this challenge may be overcome.

In conclusion, the broad spectrum of thyroid disease severity—from subclinical hypothyroidism to myxedema coma, subclinical thyrotoxicosis to thyroid storm, and microscopic papillary to anaplastic cancers—has always demanded that clinicians individualize their management of thyroid patients. Deepening knowledge of thyroid pathophysiology along with advances in diagnostic, prognostic, and therapeutic technologies applicable to thyroid diseases position this field to ride the wave of precision medicine in the decade ahead.

Acknowledgments

The author is grateful to Dr Gilbert H. Daniels and the late Dr E. Chester Ridgway, whose wise and patient mentorship nurtured his interest in thyroidology, and to the gifted mentees who inspire his confidence in its future.

Disclosure Summary: P.W.L. is a consultant to Viking Therapeutics and Veracyte.

Abbreviations

 
• CT

  computed tomography

 
• PET

  positron emission tomography.

References