- Email: [email protected]
Multiphase computed tomography for localization of parathyroid disease in patients with primary hyperparathyroidism: How many phases do we really need?
Multiphase computed tomography for localization of parathyroid disease in patients with primary hyperparathyroidism: How many phases do we really need? Salem I. Noureldine, MD,a Nafi Aygun, MD,b Michael J. Walden, MD,b Ahmed Hassoon, MD, MPH,c Sachin K. Gujar, MD,b and Ralph P. Tufano, MD, MBA, FACS,a Baltimore, MD
Background. Multiphase computed tomography (CT) involves multiple cervical CT acquisitions to accurately identify hyperfunctional parathyroid glands, thus increasing radiation exposure to the patient. We hypothesized that only 2 cervical acquisitions, instead of the conventional 4, would provide equivalent localization information and halve the radiation exposure. Methods. We identified 53 consecutive patients with primary hyperparathyroidism who underwent multiphase CT before parathyroidectomy. All scans were reinterpreted first using 2 phases then using all 4 phases. The accuracies of interpretations were determined with surgical findings serving as the standard of reference. Results. Sixty-four hyperfunctional parathyroid glands were resected with a mean weight of 394.3 mg. Twophase CT lateralized the hyperfunctional glands in 38 patients with a sensitivity, positive predictive value (PPV), and accuracy of 100%, 71.7%, and 71.7%, respectively. Four-phase CT lateralized the hyperfunctional glands in 39 patients with a sensitivity, PPV, and accuracy of 95.1%, 76.5%, and 73.6%, respectively. For quadrant localization, the accuracy of 2-phase and 4-phase CT was 50.9% and 52.8%, respectively. Conclusion. Our results suggest that 2-phase and 4-phase CT provide an equivalent diagnostic accuracy in localizing hyperfunctional parathyroid glands. The reduced radiation exposure to the patient may make 2-phase acquisitions a more acceptable alternative for preoperative localization. (Surgery 2014;156:1300-7.) From the Division of Head and Neck Endocrine Surgery, Department of Otolaryngology – Head and Neck Surgery,a the Department of Radiology and Radiological Science,b Johns Hopkins University School of Medicine, and the Department of Epidemiology,c Johns Hopkins Bloomberg School of Public Health, Baltimore, MD
HIGH-QUALITY PREOPERATIVE IMAGING is a critical component for the success of targeted parathyroidectomy in patients with primary hyperparathyroidism Presented at the American Association of Endocrine Surgeons 2014 annual meeting, April 28, 2014, Boston, Massachusetts. S.I.N. and N.A. contributed equally to this work. Financial Disclosure: The authors have no financial interests in companies or other entities that have an interest in the information included in the contribution. Conflicts of Interest: All authors report no conflicts of interest. Accepted for publication August 7, 2014. Reprint requests: Ralph P. Tufano, MD, MBA, FACS, Charles W. Cummings MD Professor, Department of Otolaryngology-Head and Neck Surgery, The Johns Hopkins School of Medicine, Johns Hopkins Outpatient Center; 601 N. Caroline Street, 6th floor Baltimore, MD 21287. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2014.08.002
1300 SURGERY
(PHPT). Currently, the majority of institutions use ultrasonography (US) and technetium-99m (Tc99m) sestamibi scanning for preoperative localization, with reported sensitivities of 50–85% and 60–92%, respectively, when deployed individually. There is a reported increase in the likelihood of successful localization of hyperfunctional parathyroid glands when US and Tc-99m sestamibi are used in combination (73-95%).1-6 Shortcomings of this approach include the inability to screen the mediastinum with US, limited spatial resolution of Tc-99m sestamibi, operator dependability, and lack of anatomic landmarks identifiable on these types of images that would help the surgeon to approach parathyroidectomy in a focused or targeted manner.7 The addition of single photon emission computed tomography (CT) and conventional CT has allowed some improvements in these limitations.1
Surgery Volume 156, Number 6
Recent studies suggest that multiphase CT, also known as 4-dimensional CT in the parathyroid imaging lexicon, has a higher sensitivity and accuracy for preoperative localization of hyperfunctional parathyroid glands in patients with PHPT than either Tc-99m sestamibi scanning, US, or both combined.1,7-10 Multiphase CT involves 4 sequential CT acquisitions through the neck once before administration of intravenous contrast material and 3 additional predetermined times after contrast is given. These separate CT acquisitions are variably referred to as precontrast, arterial, venous, and delayed phases. Multiphase CT permits visualization of tissue at 4 successive time points as the contrast bolus is passing through, which provides information about how fast and how high the contrast enhancement of various tissue is, as well as, how quickly the contrast material washes out. Because parathyroid glands exhibit different contrast enhancement and washout characteristics than lymph nodes and vessels, differentiation of hyperfunctional parathyroid glands from mimickers can be achieved.11 Multiphase CT has become a favored imaging modality, particularly for reoperative parathyroid surgery or when other preoperative imaging modalities fail to localize a hyperfunctional parathyroid gland.1,9-12 However, the main drawback of multiphase CT is its high radiation dose, with a conservative dose estimate of 27 millisievert (mSv) for a typical multiphase CT examination with 4 phases compared with 11 mSv for a Tc-99m sestamibi scan.9 Twenty-seven mSv is equivalent to the radiation dose that an average American receives from natural background radiation over 9 years.8,13 Beland et al11 have shown that hyperfunctional parathyroid glands often show rapid enhancement and early washout. Therefore, the delayed phase could be eliminated because there is no significant difference in the enhancement characteristics of lymph nodes and hyperfunctional parathyroid glands in this phase. The change from a 4-phase to a 3-phase CT protocol allowed reduction in the effective radiation dose of a multiphase CT to 21 mSv.14 We hypothesized that elimination of the precontrast and delayed phases from the typical 4-phase acquisition would not result in diminished diagnostic accuracy while halving the radiation exposure. MATERIALS AND METHODS Data source. Data were extracted from the medical record reports of all consecutive patients
Noureldine et al 1301
with biochemically confirmed PHPT (calcium level >10.4 mg/dL and/or serum intact parathyroid hormone [iPTH] level >65 pg/mL) who underwent multiphase CT for preoperative localization before undergoing primary parathyroidectomy. All multiphase CT scans were performed at our institution between September 2008 and May 2013 and interpreted by a single, board-certified neuroradiologist. All primary parathyroidectomy procedures were performed by a single highvolume surgeon at our institution between September 2008 and September 2013. Multiphase CT studies were obtained to assist in localizing hyperfunctional parathyroid glands in surgeryna€ıve patients with nonlocalizing or discordant US and sestamibi imaging, who otherwise were deemed possible candidates for targeted primary parathyroidectomy. The surgical notes were reviewed without knowledge of the 2-phase and 4phase CT readings, and the site and weight of the resected hyperfunctional parathyroid glands were recorded. This retrospective study was approved by the Institutional Review Board of Johns Hopkins Medicine, and informed consent was waived in accordance with the Health Insurance Portability and Accountability Act. Multiphase CT image processing. Imaging was performed using a Siemens (Erlangen, Germany) Somatom Definition 64 CT scanner in helical mode set from 2 cm below the carina to the skull base. The scanning parameters were 120 kVp with an effective tube current of approximately 220 mA and pitch of 0.75. Nonoverlapping 0.75 3 3.0-mm axial reconstructions were created in each of the 4 phases along with sagittal, coronal, and 2 oblique reformatted images. The first phase was without intravenous iodinated contrast and consisted of an initial unenhanced scan (precontrast phase). Twenty-five seconds before the beginning of the second phase, intravenous contrast administration of 120 mL of nonionic contrast material (Iohexol [Omnipaque 350; GE Healthcare, Little Chalfont, Buckinghamshire, United Kingdom] or Iodixanol [Visipaque 320; GE Healthcare,]) was begun at 3 mL/s (arterial phase). This timing was chosen such that imaging through the neck occurred during maximal arterial opacification. Sixty-eight seconds after the end of the second phase, the third phase (venous phase) commenced and 111 seconds after the end of the third phase, the final acquisition (delayed phase) began (Figure). The calculated mean volume CT dose index as required by the state regulatory body was recorded as the radiation absorbed dose.
1302 Noureldine et al
Surgery December 2014
Figure. Multiphase computed tomography (CT) of a patient with primary hyperparathyroidism. No lesions were identified on ultrasound or Tc-99m sestamibi. Multiphase CT demonstrates an avidly enhancing lesion (arrows) posterior to the left thyroid inferiorly with marked enhancement in the arterial phase and rapid washout of contrast in the venous phase. (A) Precontrast phase. (B) Arterial phase. (C) Venous phase. (D) Delayed phase.
Reinterpretation of multiphase CT acquisitions. Interpretation of the multiphase CT acquisitions was performed between December 2011 and November 2013. All patient-identifying information was removed from the images by a medical student who was not involved in other aspects of the study. The precontrast (phase I) and delayed (phase 4) acquisitions were extracted from the complete data sets, anonymized, and saved with unique identifiers. The remaining 2-phase datasets that included the arterial (phase 2) and venous (phase 3) phases were presented in a random fashion for interpretation. A board-certified neuroradiologist, with 10 years of post-training experience in reading head and neck CTs, independently reviewed all studies using a commercially available picture archiving and communication system (Ultravisual, Emageon, Madison, WI). The reviewer had the capability of creating multiplanar reformats and measuring attenuation coefficients as deemed necessary. The radiologist was blinded to the surgical outcomes and all clinical data, except that the subjects had PHPT with nonlocalizing or discordant US and Tc-99m sestamibi imaging results, which in our practice prompts the acquisition of a multiphase CT scan. Suspected hyperfunctional parathyroid glands were recorded with corresponding table locations. The reviewer assigned a confidence level of 1, 2, or 3 to his readings based on the location, appearance, and enhancement characteristics of the
hyperfunctional glands: 1, very confident; 3, not confident; and 2 when there was confidence in the reading, but not as strong as 1. After a period of >3 months, anonymized image sets that included all 4 phases of the CT examinations were presented to the same neuroradiologist in a random fashion and with unique identifiers that did not permit establishing a connection with the previous 2-phase CT interpretations. The locations of the suspected hyperfunctional parathyroid glands and the confidence levels for each case were recorded in a similar fashion. As per our practice, all outside Tc-99m sestamibi scans were reinterpreted preoperatively at our institution. We did not reinterpret any of the US imaging studies. The results of both imaging tests were obtained from the review of medical records. Standard of reference. The suspected locations of the hyperfunctional parathyroid glands were correlated with the surgical reports, pathology reports, and intraoperative rapid parathyroid hormone assays for each patient to determine the accuracy of each interpretation for both the 4phase and 2-phase data sets independently. Intraoperative biochemical resolution of PHPT was defined as a 50% drop from baseline and into the normal range. Because all patients had biochemical confirmation of PHPT as a condition of inclusion, the sensitivity, positive predictive value (PPV), and accuracy of the preoperative imaging tests were
Surgery Volume 156, Number 6
only calculated. A true positive was defined as the probability that the tests correctly identified the hyperfunctional parathyroid gland in cases of single adenomas or correctly identified the location of all hyperfunctional glands in patients with multigland disease. PPV was defined as the probability that a patient with positive test had a single hyperfunctional parathyroid gland or multigland disease. A false positive (FP) was defined as tests that incorrectly suggested lateralization of a hyperfunctional parathyroid gland or the presence of multigland disease, and a false negative was defined as tests incorrectly suggested the absence of hyperfunctional parathyroid glands. Lateralization accuracy was determined as the number of patients in whom the imaging study correctly identified the anatomic side of the hyperfunctional parathyroid gland. The accuracy of quadrant localization was defined as the number of patients in whom each imaging technique correctly identified all the quadrant anatomic locations (right/ left, superior/inferior) of the hyperfunctional parathyroid glands, expressed as a percentage of the total number of patients in the study. We utilized the surgical anatomic location of the identified hyperfunctional parathyroid gland for accurate quadrant localization rather than the actual embryonic origin. We believe that the anatomic location is more helpful to the surgeon in pursuing a focused or targeted parathyroidectomy (eg, superior parathyroid adenoma that has migrated caudally to lie in the inferior quadrant with respect to the thyroid lobe). Statistical analysis. Data analysis was performed using STATA version 13 (StataCorp, College Station, TX). Sensitivity, PPV, and accuracy were calculated relative to surgical findings. All values are expressed as means ± standard deviation, ranges, or absolute numbers unless specified otherwise. RESULTS A total of 53 patients were identified during the study period; 37 (69.8%) were women and 16 (30.2%) were men with a mean age at admission of 56.9 ± 11.4 years (range, 31–78). The mean preoperative serum calcium and iPTH levels were 11 ± 0.6 mg/dL (median, 11) and 95.7 ± 58.2 pg/mL (median, 82.4), respectively. Sixty-four hyperfunctional parathyroid glands were resected during surgery with a mean weight and volume of 394.3 ± 301.1 mg (median, 300) and 465.4 ± 294.1 mm3 (median, 410), respectively. At surgery, 42 patients in the study cohort (79.2%) had a single, hyperfunctional parathyroid gland.
Noureldine et al 1303
In the remaining 11 patients (20.8%), 2 hyperfunctional parathyroid glands were found. Overall, there were 14 left superior, 13 left inferior, 15 right superior, 19 right inferior, and 3 mediastinal parathyroid glands resected. US was used in all patients and Tc-99m sestamibi was used in 52 patients (98.1%) during their initial preoperative evaluation. Only 38 US (71.7%) and 42 sestamibi (80.8%) scans were performed at our institution. The remaining were requested by a referring physician from an outside institution and were not repeated at our institution. Preoperatively, all US and Tc-99m sestamibi studies were either nonlocalizing or discordant. The sensitivity, PPV, and accuracy of US was 57.1%, 52.6%, and 37.7%, respectively. The sensitivity, PPV, and accuracy of Tc-99m sestamibi imaging was 48.7%, 59.4%, and 36.5%, respectively. When combined, the sensitivity, PPV, and accuracy were 82.9%, 63.0%, and 55.8%, respectively. US and sestamibi correctly coincided in lateralizing all hyperfunctional parathyroid glands in only 9 patients (17.3%). Table I summarizes the sensitivity, PPV, and accuracy of all 4 imaging modalities in detecting the hyperfunctional parathyroid glands. Overall, 2-phase CT identified 70 potential hyperfunctional parathyroid glands and achieved a lateralization accuracy of 71.7%. Two-phase CT achieved a localization accuracy of 50.9% in correctly identifying the anatomic quadrant of all hyperfunctional parathyroid glands in 27 of the 53 patients. Four-phase CT identified 67 potential hyperfunctional parathyroid glands and achieved a lateralization accuracy of 73.6%. Four-phase CT achieved a localization accuracy of 52.8% in correctly identifying the anatomic quadrant of all hyperfunctional parathyroid glands in 28 of the 53 patients. The mean volume CT dose index for 4-phase and 2-phase CT were 63.6 mGy and 31.8 mGy, respectively. All hyperfunctional parathyroid glands identified on 2-phase and 4-phase CT were assigned a confidence level. Table II summarizes the outcome of identified hyperfunctional glands per confidence level. Of the 12 patients with FP interpretations on 4-phase CT, 50% were found to have multigland disease. In this group, the mean preoperative serum iPTH and calcium levels were 72.7 ± 31.0 pg/mL and 10.8 ± 0.7 mg/dL, respectively, and 66% of the patients had an iPTH level of <100 pg/mL. The mean weight and volume of resected glands was 313.7 mg (median, 200) and 500.9 mm3 (median, 270), respectively. In 2 patients, no hyperfunctional parathyroid glands were identified on 4-phase CT. Of those, 1 patient was found to have multigland
1304 Noureldine et al
Surgery December 2014
Table I. Sensitivity, positive predictive value, and accuracy of the 4 imaging modalities in detecting the hyperfunctional parathyroid glands Positive Patients Sensitivity predictive Accuracy Imaging modality (n) (%) value (%) (%) Four-phase CT Lateralization Localization Two-phase CT Lateralization Localization Sestamibi Lateralization Localization Ultrasound Lateralization Localization Ultrasound and sestamibi Lateralization Localization
53 39 28 53 38 27 52 19 10 53 20 9 52
— 95.1 93.3 — 100 100 — 48.7 33.3 — 57.1 37.5 —
— 76.5 54.9 — 71.7 50.9 — 59.4 31.3 — 52.6 23.7 —
— 73.6 52.8 — 71.7 50.9 — 36.5 19.2 — 37.7 17.0 —
29 14
82.9 70.0
63.0 30.4
55.8 26.9
CT, Computed tomography.
disease and the other was found to have a 180-mg solitary right inferior hyperfunctional parathyroid gland. Of the 15 patients with FP interpretations on 2-phase CT, 46.7% were found to have multigland disease. In this group, the mean preoperative serum iPTH and calcium levels were 82.3 ± 39.3 pg/mL and 10.9 ± 0.7 mg/dL, respectively, and 60% of the patients had an iPTH level of <100 pg/mL. The mean weight and volume of resected glands was 297.9 mg (median, 198) and 471.8 mm3 (median, 243), respectively. Of all the patients with FP interpretations of 4-phase CT, 2-phase CT interpretations were also FP in those patients. Of all the patients with FP interpretations of 2-phase CT, the 4-phase CT interpretations were FP in 12 patients, false negative in 1 patient, and true positive in 2 patients. When examining the 11 patients with multigland disease separately, we found that 2-phase CT localized $1 hyperfunctional parathyroid gland in all patients, whereas 4-phase CT lateralized $1 hyperfunctional gland in 10 patients and localized $1 hyperfunctional gland in 9 patients. Both 4-phase and 2-phase CT suggested the presence of multigland disease in only 4 of the 11 patients (36.4%), and both achieved a similar accuracy of 27.3% and 18.2% in lateralizing and localizing all the hyperfunctional parathyroid glands in 3 and 2 patients, respectively. Of the 3 patients with a hyperfunctional mediastinal parathyroid gland, 2-phase CT
Table II. Reader’s confidence level in lateralization of each identified hyperfunctional parathyroid gland as it correlates to surgical findings
Confidence level
Correct lateralization (%)
Incorrect identification or lateralization (%)
Reader’s confidence level of each lesion using 4-phase CT (n = 67) 1 (n = 47) 89.4 10.6 2 (n = 8) 50.0 50.0 3 (n = 12) 58.3 41.7 Reader’s confidence level of each lesion using 2-phase CT (n = 70) 1 (n = 41) 87.8 17.1 2 (n = 15) 53.3 46.7 3 (n = 14) 42.9 57.1 CT, Computed tomography.
lateralized and localized the hyperfunctional gland in all but one patient. Four-phase CT localized the hyperfunctional mediastinal gland in all 3 patients. Intraoperatively, all but 3 patients achieved biochemical resolution of the PHPT based on intraoperative rapid parathyroid hormone measurements at the conclusion of the procedure. Bilateral exploration was always performed in this scenario and 4 parathyroid glands were always accounted for. The parathyroid glands left in situ were deemed normal by the operating surgeon based on size, shape, and consistency. Two of these patients were found to be eucalcemic at the 1- and 6-month follow-up testing and 1 patient had persistent PHPT. There were no perioperative complications in this study. All patients reported no subjective change in voice postoperatively and there was no evidence of vocal fold paralysis on postoperative laryngoscopy after extubation. Per our final determinate of successful surgery, 52 patients (98.1%) were rendered eucalcemic at 6-month follow-up. The mean postoperative serum calcium level was 9.3 ± 0.6 mg/dL (median, 9.4). DISCUSSION Our study results display a mean sensitivity, PPV, and accuracy of 93.3%, 54.9%, and 52.8%, respectively, for the precise multiphase CT localization of hyperfunctional parathyroid glands when prior traditional localization studies have been unsuccessful or discordant in localizing these hyperfunctional glands. These results are concordant with those of prior studies that have shown that multiphase CT can accurately localize hyperfunctional
Noureldine et al 1305
Surgery Volume 156, Number 6
parathyroid glands not seen on US and/or Tc-99m sestamibi scans.1,11,12 Hunter et al8 recently reported an accuracy rate of 86.6% for quadrant localization of hyperfunctional parathyroid glands in patients with PHPT. Our patient cohort is different from theirs in that we only included patients with nonlocalizing or discordant US and Tc-99m sestamibi studies. Patients with concordant test results underwent targeted parathyroidectomy at our institution, because we believe this approach is more cost effective and avoids excessive radiation exposure. It is difficult to predict what impact selecting patients with nonlocalizing or discordant US or Tc99m sestamibi results will have on the accuracy of multiphase CT, although it can be speculated that this might result in a cohort with smaller glands, because it is more likely that US and Tc99m sestamibi will be concordant in cases of larger glands. The fact that the mean weight of the removed hyperfunctional parathyroid glands was 394 mg (median, 300 mg) in our study compared with 757 mg (median, 417 mg) in the Hunter et al8 study supports this hypothesis and may potentially account for the differences in reported accuracies. Also, in our study the rate of multigland disease was 20.8%. Multigland disease was the largest contributor to our FP and falsenegative results. In the Hunter et al study,8 patients with multigland disease were excluded from the localization accuracy analysis. As a tertiary referral center, we receive a lot of patients who have undergone previous localization studies and are specifically referred to us when the localization studies do not localize or are discordant, to try to reconcile these issues. Therefore, our study population was also likely skewed toward having smaller hyperfunctioning parathyroid glands. The results of our study suggest that we may eliminate the precontrast and delayed phases. It seems that there is an equivalent diagnostic accuracy in localizing hyperfunctional parathyroid glands between 2-phase and 4-phase CT scanning. It is not surprising that there was no difference in sensitivity between 2-phase and 4-phase scanning, because sensitivity is related to the radiologist’s ability to detect an abnormal anatomic feature in the neck, which is not substantially affected by the number of repeat scans. Theoretically, 4-phase compared with 2-phase CT scanning provides more detailed information about the contrast uptake and washout features of a localized parathyroid gland, which improves lesion characterization and may help to differentiate hyperfunctional glands from mimickers, thus potentially decreasing
the FP results and improving the radiologist’s confidence. These benefits did not seem to have materialized in our study; the radiologist’s confidence was considerably similar between 2-phase and 4-phase CT scans and there was no substantial difference in FP results. We believe that a change from a 4-phase to a 2-phase protocol balances the need for sufficient information from this localizing study to consider targeted parathyroidectomy against unnecessary phases and radiation exposure. Future studies should also consider comparing multiphase CT-directed surgery with bilateral exploration alone from an outcome and overall cost standpoint in this patient population. There are certain limitations within the study. The overall cohort size is relatively small to detect a meaningful proportional difference between the 2 imaging tests. This is partly because, in our practice, multiphase CT is only obtained to assist in the preoperative localization of patients with PHPT with nonlocalizing or discordant US and Tc99m sestamibi imaging. Usually, multiphase CT is not obtained as the primary localization modality. Therefore, future prospective studies, with power analysis, are warranted to better examine if 2-phase CT can really serve as an alternative to 4-phase CT for preoperative localization in patients with PHPT. In conclusion, multiphase CT can accurately localize hyperfunctional parathyroid glands not seen on US or Tc-99m sestamibi imaging in patients with PHPT. Our preliminary results demonstrate that 2-phase CT seems to provide an equivalent diagnostic accuracy to 4-phase CT in this patient population. The use of 2-phase CT would result in a reduction in the overall radiation exposure to the patient, making 2-phase CT a more acceptable alternative to 4-phase CT. REFERENCES 1. Rodgers SE, Hunter GJ, Hamberg LM, Schellingerhout D, Doherty DB, Ayers GD, et al. Improved preoperative planning for directed parathyroidectomy with 4-dimensional computed tomography. Surgery 2006;140:932-40. 2. Elaraj DM, Sippel RS, Lindsay S, Sansano I, Duh QY, Clark OH, et al. Are additional localization studies and referral indicated for patients with primary hyperparathyroidism who have negative sestamibi scan results? Arch Surg 2010; 145:578-81. 3. Siperstein A, Berber E, Barbosa GF, Tsinberg M, Greene AB, Mitchell J, et al. Predicting the success of limited exploration for primary hyperparathyroidism using ultrasound, sestamibi, and intraoperative parathyroid hormone: analysis of 1158 cases. Ann Surg 2008;248:420-8. 4. Davis ML, Quayle FJ, Middleton WD, Acosta LM, Hix-Hernandez SJ, Snyder SK, et al. Ultrasound facilitates minimally invasive parathyroidectomy in patients lacking definitive
1306 Noureldine et al
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
localization from preoperative sestamibi scan. Am J Surg 2007;194:785-90. Nichols KJ, Tomas MB, Tronco GG, Rini JN, Kunjummen BD, Heller KS, et al. Preoperative parathyroid scintigraphic lesion localization: accuracy of various types of readings. Radiology 2008;248:221-32. Udelsman R. Six hundred fifty-six consecutive explorations for primary hyperparathyroidism. Ann Surg 2002; 235:665-70. Cheung K, Wang TS, Farrokhyar F, Roman SA, Sosa JA. A meta-analysis of preoperative localization techniques for patients with primary hyperparathyroidism. Ann Surg Oncol 2012;19:577-83. Hunter GJ, Schellingerhout D, Vu TH, Perrier ND, Hamberg LM. Accuracy of four-dimensional CT for the localization of abnormal parathyroid glands in patients with primary hyperparathyroidism. Radiology 2012;264:789-95. Starker LF, Mahajan A, Bjorklund P, Sze G, Udelsman R, Carling T. 4D parathyroid CT as the initial localization study for patients with de novo primary hyperparathyroidism. Ann Surg Oncol 2011;18:1723-8. Lubitz CC, Hunter GJ, Hamberg LM, Parangi S, Ruan D, Gawande A, et al. Accuracy of 4-dimensional computed tomography in poorly localized patients with primary hyperparathyroidism. Surgery 2010;148:1129-37. Beland MD, Mayo-Smith WW, Grand DJ, Machan JT, Monchik JM. Dynamic MDCT for localization of occult parathyroid adenomas in 26 patients with primary hyperparathyroidism. AJR Am J Roentgenol 2011;196:61-5. Mortenson MM, Evans DB, Lee JE, Hunter GJ, Shellingerhout D, Vu T, et al. Parathyroid exploration in the reoperative neck: improved preoperative localization with 4D-computed tomography. J Am Coll Surg 2008;206:888-95. National Council on Radiation Protection and Measurements. Ionizing radiation exposure of the population of the United StatesNCRP Report No. 160. Bethesda (MD): National Council on Radiation Protection and Measurements; 2009. Kelly HR, Hamberg LM, Hunter GJ. 4D-CT for preoperative localization of abnormal parathyroid glands in patients with hyperparathyroidism: accuracy and ability to stratify patients by unilateral versus bilateral disease in surgery-naive and re-exploration patients. AJNR Am J Neuroradiol 2014; 35:176-81.
DISCUSSION Dr Lilah Morris (Tucson, AZ): What was the experience of your neuroradiologists in reading traditional 4-phase/4D CT before this study? Dr Salem I. Noureldine (Baltimore, MD): The neuroradiologist involved in this study, Dr Aygun, is 1 of 14 board-certified neuroradiologists at our institution who interpret 4-phase/4-dimensional (4D)CT images of the head and neck. Dr Aygun has >10 years of post-training experience in reading head and neck CT images, and has a specific interest in parathyroid disease. Before this study, Dr Aygun had initially read only 6 out of the 53 four-phase CT scans when they were ordered preoperatively. In this study, we only examined the 4-phase CT and the 2-phase CT reinterpretations. We did not intend to compare the findings of his
Surgery December 2014
initial readings (at the time the imaging study was performed) with his reinterpretations, but I can tell you that they were almost the same. Dr Lilah Morris: Can you just take the 2-phase and implement it? Or do you have to have someone who is really experienced in reading the 4-phase study? Dr Salem I. Noureldine: Of course, with time, the radiologist was more comfortable with interpreting the 4-phase CT images. We believe that, yes, especially in the community, it is probably better to initially utilize the 4-phase CT and gain more experience, before shifting to a 2-phase CT protocol. However, we did not examine the learning curve associated with these imaging techniques in this current study. Dr Wen Shen (San Francisco, CA): How many of these were first-time versus redo patients? Dr Salem I. Noureldine: These were all surgeryna€ıve patients, with nonlocalizing or discordant ultrasound and sestamibi imaging, who otherwise were deemed possible candidates for targeted primary parathyroidectomy. Dr Wen Shen: So to be provocative here, how many phases do we need? I am going to ask you, you are trying to limit radiation. Do you routinely get a CT scan in all your image-negative patients, or would you proceed if it is sestamibi negative, ultrasound negative, proceed with exploration? Do we really need the CT scan because it did not help in every case? Dr Salem I. Noureldine: That is a very good point, and I thank you for bringing it up. We believe that outcome correlates most significantly with surgical experience. As Dr Doppman once mentioned, ‘‘the best localization study is to find an experienced parathyroid surgeon.’’ However, in this scenario, when ultrasound and sestamibi were discordant or did not contribute to the localization of the parathyroid lesions, we found it more prudent to go ahead and order a 4-phase CT as the next step, rather than repeating the ultrasound and sestamibi that were performed elsewhere, with the intent to consider a targeted parathyroidectomy if possible. When 4-phase CT localized the lesion, targeted parathyroidectomy was performed with intraoperative parathyroid hormone assessment and was successful in achieving targeted resection in the majority of cases. This translates into less operating room time, less risk of complications, decreased duration of hospital stay, and an improved cosmetic result as reported in the literature. You might argue that Dr Tufano is an experienced parathyroid surgeon, and should not have a higher risk of complications with bilateral
Surgery Volume 156, Number 6
exploration while achieving similar success rates to targeted parathyroidectomy. However, most parathyroidectomies are being performed in the community by low-volume and less experienced surgeons whose complication rates are likely higher. Obtaining a multiphase CT examination would potentially assist these surgeons because they can easily interpret the CT images, and can potentially decrease their complication rates if targeted parathyroidectomy is achieved. Dr Philip Smith (Charlottesville, VA): No disclosures. More a comment in relation to the prior question; we did a similar approach at our institution. We had 5 neuroradiologists, most of whom did not have extensive experience in 4D CT. We found very similar results, that 2-phase had
Noureldine et al 1307
equivalent findings across and there is very good interrater reliability for that. So I very much agree with your findings. Dr Salem I. Noureldine: Thank you for your comment. At our institution, we also compared head and neck 4-phase CT and 2-phase CT interpretations for 2 board-certified neuroradiologists, with 10 and 7 years of post-training experience in reading neck CTs. For both, 2 phase CT had equivalent findings to 4-phase CT in localizing parathyroid lesions. When examining the kappa coefficients to measure interrater and intrarater reliability, we found that the agreement was high between both readers. There was a fair agreement between the confidences reported for each lesion and between both of the reader’s findings.
Background. Multiphase computed tomography (CT) involves multiple cervical CT acquisitions to accurately identify hyperfunctional parathyroid glands, thus increasing radiation exposure to the patient. We hypothesized that only 2 cervical acquisitions, instead of the conventional 4, would provide equivalent localization information and halve the radiation exposure. Methods. We identified 53 consecutive patients with primary hyperparathyroidism who underwent multiphase CT before parathyroidectomy. All scans were reinterpreted first using 2 phases then using all 4 phases. The accuracies of interpretations were determined with surgical findings serving as the standard of reference. Results. Sixty-four hyperfunctional parathyroid glands were resected with a mean weight of 394.3 mg. Twophase CT lateralized the hyperfunctional glands in 38 patients with a sensitivity, positive predictive value (PPV), and accuracy of 100%, 71.7%, and 71.7%, respectively. Four-phase CT lateralized the hyperfunctional glands in 39 patients with a sensitivity, PPV, and accuracy of 95.1%, 76.5%, and 73.6%, respectively. For quadrant localization, the accuracy of 2-phase and 4-phase CT was 50.9% and 52.8%, respectively. Conclusion. Our results suggest that 2-phase and 4-phase CT provide an equivalent diagnostic accuracy in localizing hyperfunctional parathyroid glands. The reduced radiation exposure to the patient may make 2-phase acquisitions a more acceptable alternative for preoperative localization. (Surgery 2014;156:1300-7.) From the Division of Head and Neck Endocrine Surgery, Department of Otolaryngology – Head and Neck Surgery,a the Department of Radiology and Radiological Science,b Johns Hopkins University School of Medicine, and the Department of Epidemiology,c Johns Hopkins Bloomberg School of Public Health, Baltimore, MD
HIGH-QUALITY PREOPERATIVE IMAGING is a critical component for the success of targeted parathyroidectomy in patients with primary hyperparathyroidism Presented at the American Association of Endocrine Surgeons 2014 annual meeting, April 28, 2014, Boston, Massachusetts. S.I.N. and N.A. contributed equally to this work. Financial Disclosure: The authors have no financial interests in companies or other entities that have an interest in the information included in the contribution. Conflicts of Interest: All authors report no conflicts of interest. Accepted for publication August 7, 2014. Reprint requests: Ralph P. Tufano, MD, MBA, FACS, Charles W. Cummings MD Professor, Department of Otolaryngology-Head and Neck Surgery, The Johns Hopkins School of Medicine, Johns Hopkins Outpatient Center; 601 N. Caroline Street, 6th floor Baltimore, MD 21287. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2014.08.002
1300 SURGERY
(PHPT). Currently, the majority of institutions use ultrasonography (US) and technetium-99m (Tc99m) sestamibi scanning for preoperative localization, with reported sensitivities of 50–85% and 60–92%, respectively, when deployed individually. There is a reported increase in the likelihood of successful localization of hyperfunctional parathyroid glands when US and Tc-99m sestamibi are used in combination (73-95%).1-6 Shortcomings of this approach include the inability to screen the mediastinum with US, limited spatial resolution of Tc-99m sestamibi, operator dependability, and lack of anatomic landmarks identifiable on these types of images that would help the surgeon to approach parathyroidectomy in a focused or targeted manner.7 The addition of single photon emission computed tomography (CT) and conventional CT has allowed some improvements in these limitations.1
Surgery Volume 156, Number 6
Recent studies suggest that multiphase CT, also known as 4-dimensional CT in the parathyroid imaging lexicon, has a higher sensitivity and accuracy for preoperative localization of hyperfunctional parathyroid glands in patients with PHPT than either Tc-99m sestamibi scanning, US, or both combined.1,7-10 Multiphase CT involves 4 sequential CT acquisitions through the neck once before administration of intravenous contrast material and 3 additional predetermined times after contrast is given. These separate CT acquisitions are variably referred to as precontrast, arterial, venous, and delayed phases. Multiphase CT permits visualization of tissue at 4 successive time points as the contrast bolus is passing through, which provides information about how fast and how high the contrast enhancement of various tissue is, as well as, how quickly the contrast material washes out. Because parathyroid glands exhibit different contrast enhancement and washout characteristics than lymph nodes and vessels, differentiation of hyperfunctional parathyroid glands from mimickers can be achieved.11 Multiphase CT has become a favored imaging modality, particularly for reoperative parathyroid surgery or when other preoperative imaging modalities fail to localize a hyperfunctional parathyroid gland.1,9-12 However, the main drawback of multiphase CT is its high radiation dose, with a conservative dose estimate of 27 millisievert (mSv) for a typical multiphase CT examination with 4 phases compared with 11 mSv for a Tc-99m sestamibi scan.9 Twenty-seven mSv is equivalent to the radiation dose that an average American receives from natural background radiation over 9 years.8,13 Beland et al11 have shown that hyperfunctional parathyroid glands often show rapid enhancement and early washout. Therefore, the delayed phase could be eliminated because there is no significant difference in the enhancement characteristics of lymph nodes and hyperfunctional parathyroid glands in this phase. The change from a 4-phase to a 3-phase CT protocol allowed reduction in the effective radiation dose of a multiphase CT to 21 mSv.14 We hypothesized that elimination of the precontrast and delayed phases from the typical 4-phase acquisition would not result in diminished diagnostic accuracy while halving the radiation exposure. MATERIALS AND METHODS Data source. Data were extracted from the medical record reports of all consecutive patients
Noureldine et al 1301
with biochemically confirmed PHPT (calcium level >10.4 mg/dL and/or serum intact parathyroid hormone [iPTH] level >65 pg/mL) who underwent multiphase CT for preoperative localization before undergoing primary parathyroidectomy. All multiphase CT scans were performed at our institution between September 2008 and May 2013 and interpreted by a single, board-certified neuroradiologist. All primary parathyroidectomy procedures were performed by a single highvolume surgeon at our institution between September 2008 and September 2013. Multiphase CT studies were obtained to assist in localizing hyperfunctional parathyroid glands in surgeryna€ıve patients with nonlocalizing or discordant US and sestamibi imaging, who otherwise were deemed possible candidates for targeted primary parathyroidectomy. The surgical notes were reviewed without knowledge of the 2-phase and 4phase CT readings, and the site and weight of the resected hyperfunctional parathyroid glands were recorded. This retrospective study was approved by the Institutional Review Board of Johns Hopkins Medicine, and informed consent was waived in accordance with the Health Insurance Portability and Accountability Act. Multiphase CT image processing. Imaging was performed using a Siemens (Erlangen, Germany) Somatom Definition 64 CT scanner in helical mode set from 2 cm below the carina to the skull base. The scanning parameters were 120 kVp with an effective tube current of approximately 220 mA and pitch of 0.75. Nonoverlapping 0.75 3 3.0-mm axial reconstructions were created in each of the 4 phases along with sagittal, coronal, and 2 oblique reformatted images. The first phase was without intravenous iodinated contrast and consisted of an initial unenhanced scan (precontrast phase). Twenty-five seconds before the beginning of the second phase, intravenous contrast administration of 120 mL of nonionic contrast material (Iohexol [Omnipaque 350; GE Healthcare, Little Chalfont, Buckinghamshire, United Kingdom] or Iodixanol [Visipaque 320; GE Healthcare,]) was begun at 3 mL/s (arterial phase). This timing was chosen such that imaging through the neck occurred during maximal arterial opacification. Sixty-eight seconds after the end of the second phase, the third phase (venous phase) commenced and 111 seconds after the end of the third phase, the final acquisition (delayed phase) began (Figure). The calculated mean volume CT dose index as required by the state regulatory body was recorded as the radiation absorbed dose.
1302 Noureldine et al
Surgery December 2014
Figure. Multiphase computed tomography (CT) of a patient with primary hyperparathyroidism. No lesions were identified on ultrasound or Tc-99m sestamibi. Multiphase CT demonstrates an avidly enhancing lesion (arrows) posterior to the left thyroid inferiorly with marked enhancement in the arterial phase and rapid washout of contrast in the venous phase. (A) Precontrast phase. (B) Arterial phase. (C) Venous phase. (D) Delayed phase.
Reinterpretation of multiphase CT acquisitions. Interpretation of the multiphase CT acquisitions was performed between December 2011 and November 2013. All patient-identifying information was removed from the images by a medical student who was not involved in other aspects of the study. The precontrast (phase I) and delayed (phase 4) acquisitions were extracted from the complete data sets, anonymized, and saved with unique identifiers. The remaining 2-phase datasets that included the arterial (phase 2) and venous (phase 3) phases were presented in a random fashion for interpretation. A board-certified neuroradiologist, with 10 years of post-training experience in reading head and neck CTs, independently reviewed all studies using a commercially available picture archiving and communication system (Ultravisual, Emageon, Madison, WI). The reviewer had the capability of creating multiplanar reformats and measuring attenuation coefficients as deemed necessary. The radiologist was blinded to the surgical outcomes and all clinical data, except that the subjects had PHPT with nonlocalizing or discordant US and Tc-99m sestamibi imaging results, which in our practice prompts the acquisition of a multiphase CT scan. Suspected hyperfunctional parathyroid glands were recorded with corresponding table locations. The reviewer assigned a confidence level of 1, 2, or 3 to his readings based on the location, appearance, and enhancement characteristics of the
hyperfunctional glands: 1, very confident; 3, not confident; and 2 when there was confidence in the reading, but not as strong as 1. After a period of >3 months, anonymized image sets that included all 4 phases of the CT examinations were presented to the same neuroradiologist in a random fashion and with unique identifiers that did not permit establishing a connection with the previous 2-phase CT interpretations. The locations of the suspected hyperfunctional parathyroid glands and the confidence levels for each case were recorded in a similar fashion. As per our practice, all outside Tc-99m sestamibi scans were reinterpreted preoperatively at our institution. We did not reinterpret any of the US imaging studies. The results of both imaging tests were obtained from the review of medical records. Standard of reference. The suspected locations of the hyperfunctional parathyroid glands were correlated with the surgical reports, pathology reports, and intraoperative rapid parathyroid hormone assays for each patient to determine the accuracy of each interpretation for both the 4phase and 2-phase data sets independently. Intraoperative biochemical resolution of PHPT was defined as a 50% drop from baseline and into the normal range. Because all patients had biochemical confirmation of PHPT as a condition of inclusion, the sensitivity, positive predictive value (PPV), and accuracy of the preoperative imaging tests were
Surgery Volume 156, Number 6
only calculated. A true positive was defined as the probability that the tests correctly identified the hyperfunctional parathyroid gland in cases of single adenomas or correctly identified the location of all hyperfunctional glands in patients with multigland disease. PPV was defined as the probability that a patient with positive test had a single hyperfunctional parathyroid gland or multigland disease. A false positive (FP) was defined as tests that incorrectly suggested lateralization of a hyperfunctional parathyroid gland or the presence of multigland disease, and a false negative was defined as tests incorrectly suggested the absence of hyperfunctional parathyroid glands. Lateralization accuracy was determined as the number of patients in whom the imaging study correctly identified the anatomic side of the hyperfunctional parathyroid gland. The accuracy of quadrant localization was defined as the number of patients in whom each imaging technique correctly identified all the quadrant anatomic locations (right/ left, superior/inferior) of the hyperfunctional parathyroid glands, expressed as a percentage of the total number of patients in the study. We utilized the surgical anatomic location of the identified hyperfunctional parathyroid gland for accurate quadrant localization rather than the actual embryonic origin. We believe that the anatomic location is more helpful to the surgeon in pursuing a focused or targeted parathyroidectomy (eg, superior parathyroid adenoma that has migrated caudally to lie in the inferior quadrant with respect to the thyroid lobe). Statistical analysis. Data analysis was performed using STATA version 13 (StataCorp, College Station, TX). Sensitivity, PPV, and accuracy were calculated relative to surgical findings. All values are expressed as means ± standard deviation, ranges, or absolute numbers unless specified otherwise. RESULTS A total of 53 patients were identified during the study period; 37 (69.8%) were women and 16 (30.2%) were men with a mean age at admission of 56.9 ± 11.4 years (range, 31–78). The mean preoperative serum calcium and iPTH levels were 11 ± 0.6 mg/dL (median, 11) and 95.7 ± 58.2 pg/mL (median, 82.4), respectively. Sixty-four hyperfunctional parathyroid glands were resected during surgery with a mean weight and volume of 394.3 ± 301.1 mg (median, 300) and 465.4 ± 294.1 mm3 (median, 410), respectively. At surgery, 42 patients in the study cohort (79.2%) had a single, hyperfunctional parathyroid gland.
Noureldine et al 1303
In the remaining 11 patients (20.8%), 2 hyperfunctional parathyroid glands were found. Overall, there were 14 left superior, 13 left inferior, 15 right superior, 19 right inferior, and 3 mediastinal parathyroid glands resected. US was used in all patients and Tc-99m sestamibi was used in 52 patients (98.1%) during their initial preoperative evaluation. Only 38 US (71.7%) and 42 sestamibi (80.8%) scans were performed at our institution. The remaining were requested by a referring physician from an outside institution and were not repeated at our institution. Preoperatively, all US and Tc-99m sestamibi studies were either nonlocalizing or discordant. The sensitivity, PPV, and accuracy of US was 57.1%, 52.6%, and 37.7%, respectively. The sensitivity, PPV, and accuracy of Tc-99m sestamibi imaging was 48.7%, 59.4%, and 36.5%, respectively. When combined, the sensitivity, PPV, and accuracy were 82.9%, 63.0%, and 55.8%, respectively. US and sestamibi correctly coincided in lateralizing all hyperfunctional parathyroid glands in only 9 patients (17.3%). Table I summarizes the sensitivity, PPV, and accuracy of all 4 imaging modalities in detecting the hyperfunctional parathyroid glands. Overall, 2-phase CT identified 70 potential hyperfunctional parathyroid glands and achieved a lateralization accuracy of 71.7%. Two-phase CT achieved a localization accuracy of 50.9% in correctly identifying the anatomic quadrant of all hyperfunctional parathyroid glands in 27 of the 53 patients. Four-phase CT identified 67 potential hyperfunctional parathyroid glands and achieved a lateralization accuracy of 73.6%. Four-phase CT achieved a localization accuracy of 52.8% in correctly identifying the anatomic quadrant of all hyperfunctional parathyroid glands in 28 of the 53 patients. The mean volume CT dose index for 4-phase and 2-phase CT were 63.6 mGy and 31.8 mGy, respectively. All hyperfunctional parathyroid glands identified on 2-phase and 4-phase CT were assigned a confidence level. Table II summarizes the outcome of identified hyperfunctional glands per confidence level. Of the 12 patients with FP interpretations on 4-phase CT, 50% were found to have multigland disease. In this group, the mean preoperative serum iPTH and calcium levels were 72.7 ± 31.0 pg/mL and 10.8 ± 0.7 mg/dL, respectively, and 66% of the patients had an iPTH level of <100 pg/mL. The mean weight and volume of resected glands was 313.7 mg (median, 200) and 500.9 mm3 (median, 270), respectively. In 2 patients, no hyperfunctional parathyroid glands were identified on 4-phase CT. Of those, 1 patient was found to have multigland
1304 Noureldine et al
Surgery December 2014
Table I. Sensitivity, positive predictive value, and accuracy of the 4 imaging modalities in detecting the hyperfunctional parathyroid glands Positive Patients Sensitivity predictive Accuracy Imaging modality (n) (%) value (%) (%) Four-phase CT Lateralization Localization Two-phase CT Lateralization Localization Sestamibi Lateralization Localization Ultrasound Lateralization Localization Ultrasound and sestamibi Lateralization Localization
53 39 28 53 38 27 52 19 10 53 20 9 52
— 95.1 93.3 — 100 100 — 48.7 33.3 — 57.1 37.5 —
— 76.5 54.9 — 71.7 50.9 — 59.4 31.3 — 52.6 23.7 —
— 73.6 52.8 — 71.7 50.9 — 36.5 19.2 — 37.7 17.0 —
29 14
82.9 70.0
63.0 30.4
55.8 26.9
CT, Computed tomography.
disease and the other was found to have a 180-mg solitary right inferior hyperfunctional parathyroid gland. Of the 15 patients with FP interpretations on 2-phase CT, 46.7% were found to have multigland disease. In this group, the mean preoperative serum iPTH and calcium levels were 82.3 ± 39.3 pg/mL and 10.9 ± 0.7 mg/dL, respectively, and 60% of the patients had an iPTH level of <100 pg/mL. The mean weight and volume of resected glands was 297.9 mg (median, 198) and 471.8 mm3 (median, 243), respectively. Of all the patients with FP interpretations of 4-phase CT, 2-phase CT interpretations were also FP in those patients. Of all the patients with FP interpretations of 2-phase CT, the 4-phase CT interpretations were FP in 12 patients, false negative in 1 patient, and true positive in 2 patients. When examining the 11 patients with multigland disease separately, we found that 2-phase CT localized $1 hyperfunctional parathyroid gland in all patients, whereas 4-phase CT lateralized $1 hyperfunctional gland in 10 patients and localized $1 hyperfunctional gland in 9 patients. Both 4-phase and 2-phase CT suggested the presence of multigland disease in only 4 of the 11 patients (36.4%), and both achieved a similar accuracy of 27.3% and 18.2% in lateralizing and localizing all the hyperfunctional parathyroid glands in 3 and 2 patients, respectively. Of the 3 patients with a hyperfunctional mediastinal parathyroid gland, 2-phase CT
Table II. Reader’s confidence level in lateralization of each identified hyperfunctional parathyroid gland as it correlates to surgical findings
Confidence level
Correct lateralization (%)
Incorrect identification or lateralization (%)
Reader’s confidence level of each lesion using 4-phase CT (n = 67) 1 (n = 47) 89.4 10.6 2 (n = 8) 50.0 50.0 3 (n = 12) 58.3 41.7 Reader’s confidence level of each lesion using 2-phase CT (n = 70) 1 (n = 41) 87.8 17.1 2 (n = 15) 53.3 46.7 3 (n = 14) 42.9 57.1 CT, Computed tomography.
lateralized and localized the hyperfunctional gland in all but one patient. Four-phase CT localized the hyperfunctional mediastinal gland in all 3 patients. Intraoperatively, all but 3 patients achieved biochemical resolution of the PHPT based on intraoperative rapid parathyroid hormone measurements at the conclusion of the procedure. Bilateral exploration was always performed in this scenario and 4 parathyroid glands were always accounted for. The parathyroid glands left in situ were deemed normal by the operating surgeon based on size, shape, and consistency. Two of these patients were found to be eucalcemic at the 1- and 6-month follow-up testing and 1 patient had persistent PHPT. There were no perioperative complications in this study. All patients reported no subjective change in voice postoperatively and there was no evidence of vocal fold paralysis on postoperative laryngoscopy after extubation. Per our final determinate of successful surgery, 52 patients (98.1%) were rendered eucalcemic at 6-month follow-up. The mean postoperative serum calcium level was 9.3 ± 0.6 mg/dL (median, 9.4). DISCUSSION Our study results display a mean sensitivity, PPV, and accuracy of 93.3%, 54.9%, and 52.8%, respectively, for the precise multiphase CT localization of hyperfunctional parathyroid glands when prior traditional localization studies have been unsuccessful or discordant in localizing these hyperfunctional glands. These results are concordant with those of prior studies that have shown that multiphase CT can accurately localize hyperfunctional
Noureldine et al 1305
Surgery Volume 156, Number 6
parathyroid glands not seen on US and/or Tc-99m sestamibi scans.1,11,12 Hunter et al8 recently reported an accuracy rate of 86.6% for quadrant localization of hyperfunctional parathyroid glands in patients with PHPT. Our patient cohort is different from theirs in that we only included patients with nonlocalizing or discordant US and Tc-99m sestamibi studies. Patients with concordant test results underwent targeted parathyroidectomy at our institution, because we believe this approach is more cost effective and avoids excessive radiation exposure. It is difficult to predict what impact selecting patients with nonlocalizing or discordant US or Tc99m sestamibi results will have on the accuracy of multiphase CT, although it can be speculated that this might result in a cohort with smaller glands, because it is more likely that US and Tc99m sestamibi will be concordant in cases of larger glands. The fact that the mean weight of the removed hyperfunctional parathyroid glands was 394 mg (median, 300 mg) in our study compared with 757 mg (median, 417 mg) in the Hunter et al8 study supports this hypothesis and may potentially account for the differences in reported accuracies. Also, in our study the rate of multigland disease was 20.8%. Multigland disease was the largest contributor to our FP and falsenegative results. In the Hunter et al study,8 patients with multigland disease were excluded from the localization accuracy analysis. As a tertiary referral center, we receive a lot of patients who have undergone previous localization studies and are specifically referred to us when the localization studies do not localize or are discordant, to try to reconcile these issues. Therefore, our study population was also likely skewed toward having smaller hyperfunctioning parathyroid glands. The results of our study suggest that we may eliminate the precontrast and delayed phases. It seems that there is an equivalent diagnostic accuracy in localizing hyperfunctional parathyroid glands between 2-phase and 4-phase CT scanning. It is not surprising that there was no difference in sensitivity between 2-phase and 4-phase scanning, because sensitivity is related to the radiologist’s ability to detect an abnormal anatomic feature in the neck, which is not substantially affected by the number of repeat scans. Theoretically, 4-phase compared with 2-phase CT scanning provides more detailed information about the contrast uptake and washout features of a localized parathyroid gland, which improves lesion characterization and may help to differentiate hyperfunctional glands from mimickers, thus potentially decreasing
the FP results and improving the radiologist’s confidence. These benefits did not seem to have materialized in our study; the radiologist’s confidence was considerably similar between 2-phase and 4-phase CT scans and there was no substantial difference in FP results. We believe that a change from a 4-phase to a 2-phase protocol balances the need for sufficient information from this localizing study to consider targeted parathyroidectomy against unnecessary phases and radiation exposure. Future studies should also consider comparing multiphase CT-directed surgery with bilateral exploration alone from an outcome and overall cost standpoint in this patient population. There are certain limitations within the study. The overall cohort size is relatively small to detect a meaningful proportional difference between the 2 imaging tests. This is partly because, in our practice, multiphase CT is only obtained to assist in the preoperative localization of patients with PHPT with nonlocalizing or discordant US and Tc99m sestamibi imaging. Usually, multiphase CT is not obtained as the primary localization modality. Therefore, future prospective studies, with power analysis, are warranted to better examine if 2-phase CT can really serve as an alternative to 4-phase CT for preoperative localization in patients with PHPT. In conclusion, multiphase CT can accurately localize hyperfunctional parathyroid glands not seen on US or Tc-99m sestamibi imaging in patients with PHPT. Our preliminary results demonstrate that 2-phase CT seems to provide an equivalent diagnostic accuracy to 4-phase CT in this patient population. The use of 2-phase CT would result in a reduction in the overall radiation exposure to the patient, making 2-phase CT a more acceptable alternative to 4-phase CT. REFERENCES 1. Rodgers SE, Hunter GJ, Hamberg LM, Schellingerhout D, Doherty DB, Ayers GD, et al. Improved preoperative planning for directed parathyroidectomy with 4-dimensional computed tomography. Surgery 2006;140:932-40. 2. Elaraj DM, Sippel RS, Lindsay S, Sansano I, Duh QY, Clark OH, et al. Are additional localization studies and referral indicated for patients with primary hyperparathyroidism who have negative sestamibi scan results? Arch Surg 2010; 145:578-81. 3. Siperstein A, Berber E, Barbosa GF, Tsinberg M, Greene AB, Mitchell J, et al. Predicting the success of limited exploration for primary hyperparathyroidism using ultrasound, sestamibi, and intraoperative parathyroid hormone: analysis of 1158 cases. Ann Surg 2008;248:420-8. 4. Davis ML, Quayle FJ, Middleton WD, Acosta LM, Hix-Hernandez SJ, Snyder SK, et al. Ultrasound facilitates minimally invasive parathyroidectomy in patients lacking definitive
1306 Noureldine et al
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
localization from preoperative sestamibi scan. Am J Surg 2007;194:785-90. Nichols KJ, Tomas MB, Tronco GG, Rini JN, Kunjummen BD, Heller KS, et al. Preoperative parathyroid scintigraphic lesion localization: accuracy of various types of readings. Radiology 2008;248:221-32. Udelsman R. Six hundred fifty-six consecutive explorations for primary hyperparathyroidism. Ann Surg 2002; 235:665-70. Cheung K, Wang TS, Farrokhyar F, Roman SA, Sosa JA. A meta-analysis of preoperative localization techniques for patients with primary hyperparathyroidism. Ann Surg Oncol 2012;19:577-83. Hunter GJ, Schellingerhout D, Vu TH, Perrier ND, Hamberg LM. Accuracy of four-dimensional CT for the localization of abnormal parathyroid glands in patients with primary hyperparathyroidism. Radiology 2012;264:789-95. Starker LF, Mahajan A, Bjorklund P, Sze G, Udelsman R, Carling T. 4D parathyroid CT as the initial localization study for patients with de novo primary hyperparathyroidism. Ann Surg Oncol 2011;18:1723-8. Lubitz CC, Hunter GJ, Hamberg LM, Parangi S, Ruan D, Gawande A, et al. Accuracy of 4-dimensional computed tomography in poorly localized patients with primary hyperparathyroidism. Surgery 2010;148:1129-37. Beland MD, Mayo-Smith WW, Grand DJ, Machan JT, Monchik JM. Dynamic MDCT for localization of occult parathyroid adenomas in 26 patients with primary hyperparathyroidism. AJR Am J Roentgenol 2011;196:61-5. Mortenson MM, Evans DB, Lee JE, Hunter GJ, Shellingerhout D, Vu T, et al. Parathyroid exploration in the reoperative neck: improved preoperative localization with 4D-computed tomography. J Am Coll Surg 2008;206:888-95. National Council on Radiation Protection and Measurements. Ionizing radiation exposure of the population of the United StatesNCRP Report No. 160. Bethesda (MD): National Council on Radiation Protection and Measurements; 2009. Kelly HR, Hamberg LM, Hunter GJ. 4D-CT for preoperative localization of abnormal parathyroid glands in patients with hyperparathyroidism: accuracy and ability to stratify patients by unilateral versus bilateral disease in surgery-naive and re-exploration patients. AJNR Am J Neuroradiol 2014; 35:176-81.
DISCUSSION Dr Lilah Morris (Tucson, AZ): What was the experience of your neuroradiologists in reading traditional 4-phase/4D CT before this study? Dr Salem I. Noureldine (Baltimore, MD): The neuroradiologist involved in this study, Dr Aygun, is 1 of 14 board-certified neuroradiologists at our institution who interpret 4-phase/4-dimensional (4D)CT images of the head and neck. Dr Aygun has >10 years of post-training experience in reading head and neck CT images, and has a specific interest in parathyroid disease. Before this study, Dr Aygun had initially read only 6 out of the 53 four-phase CT scans when they were ordered preoperatively. In this study, we only examined the 4-phase CT and the 2-phase CT reinterpretations. We did not intend to compare the findings of his
Surgery December 2014
initial readings (at the time the imaging study was performed) with his reinterpretations, but I can tell you that they were almost the same. Dr Lilah Morris: Can you just take the 2-phase and implement it? Or do you have to have someone who is really experienced in reading the 4-phase study? Dr Salem I. Noureldine: Of course, with time, the radiologist was more comfortable with interpreting the 4-phase CT images. We believe that, yes, especially in the community, it is probably better to initially utilize the 4-phase CT and gain more experience, before shifting to a 2-phase CT protocol. However, we did not examine the learning curve associated with these imaging techniques in this current study. Dr Wen Shen (San Francisco, CA): How many of these were first-time versus redo patients? Dr Salem I. Noureldine: These were all surgeryna€ıve patients, with nonlocalizing or discordant ultrasound and sestamibi imaging, who otherwise were deemed possible candidates for targeted primary parathyroidectomy. Dr Wen Shen: So to be provocative here, how many phases do we need? I am going to ask you, you are trying to limit radiation. Do you routinely get a CT scan in all your image-negative patients, or would you proceed if it is sestamibi negative, ultrasound negative, proceed with exploration? Do we really need the CT scan because it did not help in every case? Dr Salem I. Noureldine: That is a very good point, and I thank you for bringing it up. We believe that outcome correlates most significantly with surgical experience. As Dr Doppman once mentioned, ‘‘the best localization study is to find an experienced parathyroid surgeon.’’ However, in this scenario, when ultrasound and sestamibi were discordant or did not contribute to the localization of the parathyroid lesions, we found it more prudent to go ahead and order a 4-phase CT as the next step, rather than repeating the ultrasound and sestamibi that were performed elsewhere, with the intent to consider a targeted parathyroidectomy if possible. When 4-phase CT localized the lesion, targeted parathyroidectomy was performed with intraoperative parathyroid hormone assessment and was successful in achieving targeted resection in the majority of cases. This translates into less operating room time, less risk of complications, decreased duration of hospital stay, and an improved cosmetic result as reported in the literature. You might argue that Dr Tufano is an experienced parathyroid surgeon, and should not have a higher risk of complications with bilateral
Surgery Volume 156, Number 6
exploration while achieving similar success rates to targeted parathyroidectomy. However, most parathyroidectomies are being performed in the community by low-volume and less experienced surgeons whose complication rates are likely higher. Obtaining a multiphase CT examination would potentially assist these surgeons because they can easily interpret the CT images, and can potentially decrease their complication rates if targeted parathyroidectomy is achieved. Dr Philip Smith (Charlottesville, VA): No disclosures. More a comment in relation to the prior question; we did a similar approach at our institution. We had 5 neuroradiologists, most of whom did not have extensive experience in 4D CT. We found very similar results, that 2-phase had
Noureldine et al 1307
equivalent findings across and there is very good interrater reliability for that. So I very much agree with your findings. Dr Salem I. Noureldine: Thank you for your comment. At our institution, we also compared head and neck 4-phase CT and 2-phase CT interpretations for 2 board-certified neuroradiologists, with 10 and 7 years of post-training experience in reading neck CTs. For both, 2 phase CT had equivalent findings to 4-phase CT in localizing parathyroid lesions. When examining the kappa coefficients to measure interrater and intrarater reliability, we found that the agreement was high between both readers. There was a fair agreement between the confidences reported for each lesion and between both of the reader’s findings.