George J. Hunter

Hyperparathyroidism became unequivocally recognized and accepted as the cause of the debilitating bone disease osteitis fibrosa cystica circa 1925. It was at that time that Edward Churchill and Oliver Cope successfully removed a mediastinal parathyroid adenoma at the Massachusetts General Hospital in Boston (MGH). Subsequently, parathyroid surgery slowly developed around the world. By 1950, surgeons at the Mayo Clinic had collected a series of 140 cases, while MGH had accumulated 230 cases. At that time the conventional treatment of hyperparathyroidism was a whole neck exploration through a central collar incision, with a cure rate of around 95% in experienced hands.

Arteriographic localization of parathyroid lesions began in the mid-1950s. However, it was not until the late 1960s that selective venous sampling and selective arteriography became routine. Also at that time, noninvasive imaging studies, initially using selenomethionine, and later, subtraction thallium imaging, came into use for localization preoperatively. In the late 1970s the new modality of CT was investigated for the identification of parathyroid lesions. However, it was not routinely adopted, as the capabilities of the technology were limited and a good surgeon could locate abnormal glands better than any of the available imaging techniques. Technetium Tc99m sestamibi nuclear medicine scans replaced thallium imaging in the late 1980s as the method of choice for preoperative localization of parathyroid lesions, and venous sampling slowly became an uncommon procedure. Ultrasound equipment improved, and by the 1990s it was also routinely used for localization. Both of these techniques continue to the present time.

Despite the high success rate of whole neck exploration, the late 1990s saw the emergence of minimally invasive or targeted parathyroidectomy. This approach allowed day surgery with minimal complications. Key to the success of this approach to hyperparathyroidism was accurate identification of the offending gland prior to surgery, coupled with intraoperative parathyroid hormone assay. Accurate preoperative localization was particularly important in those cases requiring repeat surgery, for which the success rate without presurgical localization was around 80% in experienced hands.

In 2004, one of my endocrine surgical colleagues asked me if there was any way that we could improve on the accuracy of ultrasound (~60%), or sestamibi (~60%) for preoperative localization to further advance minimally invasive surgery. In concert with our medical physicist it was decided to try a functional approach to identification of abnormal parathyroid tissue and to apply a minimalistic perfusion imaging protocol to the problem. This approach leverages high-resolution, multidetector CT technology, an understanding of the pathophysiology of parathyroid disease, and a detailed knowledge of the embryologic anatomy of parathyroid glands. We designed a protocol to identify 3D anatomy coupled with timing information about the passage of contrast through tissues—hence 4D-CT. Our initial protocol had 4 phases: precontrast followed by 3 postcontrast phases at standardized time intervals.

The 4D-CT approach worked and demonstrated an accuracy of approximately 94% in patients with primary hyperparathyroidism who had not had prior surgery. In patients who were to undergo a repeat surgery, 4D-CT localization accuracy was about 82%. 4D-CT was able to reliably identify mediastinal, high ectopic, and intrathyroidal lesions. Supernumerary lesions were also identifiable. After we had imaged about 500 patients, it became clear that the fourth phase did not improve our accuracy, and we reduced the protocol to a single precontrast scan followed by timed early and late vascular postcontrast phases. Many centers around the country have deployed 4D-CT with a variety of modifications to our original protocol. In the 12 years since the first study, 4D-CT has become standard of care for those cases where other localization methods have failed to identify the lesion(s).

We have now imaged over 1700 patients, and currently, most requests for 4D-CT localization arise when ultrasound or sestamibi fails to indicate the location of lesion(s). In this particular study, we performed a retrospective review of over 200 patients in order to determine the accuracy of 4D-CT to correctly triage patients to minimally invasive parathyroidectomy versus bilateral neck exploration, including a cohort of 53 patients who had undergone prior unsuccessful parathyroid surgery. We demonstrated that 4D-CT could accurately predict unilateral or bilateral disease with an overall accuracy of 90%. In the re-exploration cohort, 4D-CT correctly identified unilateral versus bilateral disease in 96% of patients. The latter finding is of high clinical relevance, as many surgeons are reluctant to attempt re-exploration without a clear target on preoperative imaging.

Ongoing work is in the arena of automated decision support, better understanding of the parameters that allow identification of parathyroid disease versus lymph nodes or thyroid nodules, and the application of “Big Data” to help answer some of these needs.

Read this article at AJNR.org …