A, Patient 1, ambient light image. Mobilized thyroid lobe (white arrowhead); parathyroid gland (blue arrowhead). B, Patient 1, NIR autofluorescence. C, Patient 2, ambient light image. Mobilized thyroid lobe (white arrowhead); parathyroid gland (blue arrowhead). D, Patient 2, NIR autofluorescence.
A, Patient 3, ambient light image. Mobilized thyroid lobe (white arrowhead); parathyroid gland (blue arrowhead). B, Patient 3, near-infrared (NIR) image. Absence of parathyroid autofluorescence (false-negative). C, Patient 4, ambient light image. Ex vivo image of resected specimen, confirmed to be nonparathyroid fibroadipose tissue on frozen section and final pathology analysis. D, Patient 4, NIR image. Aberrant autofluorescence of nonparathyroid fibroadipose tissue (false-positive).
eFigure 1. Ambient Light Image of Resected Parathyroid Adenoma From Non-MEN1 Patient Ex Vivo
eFigure 2. NIR Autofluorescence Ex Vivo of Parathyroid Adenoma From Non-MEN1 Patient
Customize your JAMA Network experience by selecting one or more topics from the list below.
Squires MH, Shirley LA, Shen C, Jarvis R, Phay JE. Intraoperative Autofluorescence Parathyroid Identification in Patients With Multiple Endocrine Neoplasia Type 1. JAMA Otolaryngol Head Neck Surg. 2019;145(10):897–902. doi:10.1001/jamaoto.2019.1987
What are the results of near-infrared autofluorescence imaging for intraoperative parathyroid gland identification in patients with multiple endocrine neoplasia type 1 and primary hyperparathyroidism?
In this cohort study of 71 patients, the mean absolute value of in situ parathyroid autofluorescence was significantly lower for patients with vs without multiple endocrine neoplasia type 1; false-negative nonfluorescent parathyroid adenoma rates were higher among patients with vs without multiple endocrine neoplasia type 1 (50% vs 9%). Nonparathyroid fibroadipose tissue of patients with multiple endocrine neoplasia type 1 exhibited greater background autofluorescence, and was associated with high false-positive rates (83%) vs only 5% false-positive autofluorescence among patients without the syndrome.
Decreased parathyroid autofluorescence and increased background autofluorescence of nonparathyroid tissue among patients with multiple endocrine neoplasia type 1 with primary hyperparathyroidism may be associated with high rates of false-negative and false-positive autofluorescence, potentially limiting the utility of this intraoperative imaging adjunct within this patient subset.
Intrinsic near-infrared (NIR) autofluorescence of the parathyroid gland enables intraoperative gland identification without the need for contrast agent injection. However, whether real-time autofluorescence imaging is useful in patients with multiple endocrine neoplasia type 1 (MEN1) and primary hyperparathyroidism is unknown.
To compare quantified intraoperative parathyroid autofluorescence imaging results for patients with MEN1-associated vs those with non-MEN1 sporadic primary hyperparathyroidism.
Design, Setting, and Participants
A retrospective analysis of prospectively collected data on a cohort of 71 consecutive patients undergoing surgery for primary hyperparathyroidism by 2 experienced endocrine surgeons between June 1, 2017, and July 31, 2018, was conducted. Intraoperative imaging was performed with a handheld NIR autofluorescence device and images were captured for analysis. Post hoc blinded imaging analysis was conducted with Image J software to quantify representative areas of greatest autofluorescence from the parathyroid, thyroid, and adjacent soft tissue.
Main Outcomes and Measures
Primary end points were parathyroid autofluorescence and background thyroid and soft tissue autofluorescence, reported as median values with interquartile ranges. Rates of false-negative (lack of significant parathyroid gland autofluorescence compared with background autofluorescence, defined as parathyroid autofluorescence-background autofluorescence ratio <1.10) and false-positive autofluorescence (aberrant autofluorescence of nonparathyroid tissue confirmed by pathologic testing) were analyzed.
Of the 71 consecutive patients with primary hyperparathyroidism who underwent parathyroidectomy during the study period, 6 patients had genetically or clinically diagnosed MEN1 and 65 had sporadic non-MEN1 hyperparathyroidism. Most patients were women (MEN1: 4 [67%]; non-MEN1: 51 [78%]). Median (interquartile range) age was 49.0 (38.0-53.8) years in the MEN1 cohort and 61.0 (54.0-67.0) years in the non-MEN1 cohort. No clinically significant differences in serum preoperative parathyroid hormone level or parathyroid gland size or weight on pathologic examination were observed between the 2 cohorts. The median absolute value of in situ parathyroid autofluorescence was significantly lower in the MEN1 cohort than the non-MEN1 cohort (54.4 vs 74.3; Hedges g = −1.03; 95% CI, −1.89 to −0.17), as was the ratio of parathyroid to background autofluorescence (1.08 vs 1.59; g = −1.59; 95% CI, −2.23 to −0.96). Three patients (50%) with MEN1 had false-negative nonfluorescent parathyroid adenomas vs 6 patients (9%) without MEN1. Nonparathyroid fibroadipose tissue of patients with MEN1 exhibited greater background autofluorescence, leading to high false-positive rates (5 of 6 patients [83%]) vs only 3 of 65 (5%) false-positive autofluorescence nonparathyroid specimens among patients without MEN1.
Conclusions and Relevance
Intraoperative identification of parathyroid glands using their autofluorescence by real-time NIR imaging appears to have utility in patients with primary hyperparathyroidism. In this initial cohort of patients with MEN1, decreased parathyroid autofluorescence and increased background autofluorescence of nonparathyroid tissue may be associated with high rates of false-negative and false-positive fluorescence, potentially limiting the utility of this adjunct in this specific subset of patients.
Detection of intrinsic near-infrared (NIR) autofluorescence of the parathyroid gland enables intraoperative gland identification and localization without the need for contrast agent injection. Multiple imaging systems have been developed to allow real-time imaging of parathyroid glands to improve localization of parathyroid adenomas and decrease the risk of injury to healthy parathyroid glands. Several groups have published their initial experiences, using either the Fluobeam 800 Clinic Imaging Device (Fluoptics USA) or the Parathyroid Detection PTEye System (AiBioMed), both of which recently received US Food and Drug Administration approval.1-14 A recent study reported preliminary results of a prospective clinical trial examining the utility of a novel, handheld NIR autofluorescence imaging device, the PDE-NEO II (Hamamatsu, Mitaka USA, Inc), for intraoperative localization and identification of parathyroid glands among patients undergoing parathyroidectomy for a diagnosis of primary hyperparathyroidism.15 Patients with a diagnosis of multiple endocrine neoplasia (MEN) were excluded from that initial feasibility study, however, and to our knowledge, no study has yet examined parathyroid autofluorescence patterns in patients with MEN. Whether real-time autofluorescence imaging is useful for parathyroid gland localization in patients with MEN type 1 (MEN1) and primary hyperparathyroidism remains unknown.
MEN1-associated hyperparathyroidism is typically secondary to asymmetric pathologic hyperplasia of 3 or more parathyroid glands.16,17 Given the risk of recurrent disease with limited surgery, the American Association of Endocrine Surgeons guidelines recommend bilateral neck exploration and subtotal parathyroidectomy for most patients with hyperparathyroidism in the setting of MEN1.18 Although both healthy and pathologic parathyroid glands appear to have intrinsic autofluorescence due to an as-yet uncharacterized fluorophore,19,20 anecdotal reports suggest that autofluorescent localization of parathyroid glands in patients with MEN1 may be more challenging. The aim of the present study was to evaluate the utility of real-time intraoperative NIR autofluorescence imaging in patients with MEN1 undergoing surgery for primary hyperparathyroidism.
All patients undergoing surgery for a diagnosis of primary hyperparathyroidism between June 1, 2017, and July 31, 2018, at The Ohio State University were enrolled in a prospective institutional clinical study. Preliminary results of this study were recently reported, although all patients with a diagnosis of MEN1 were excluded from that initial analysis.15 A clinical diagnosis of MEN1 was made according to accepted multidisciplinary practice guidelines based on the presence of 2 or more of the 3 MEN1-related endocrine disorders, namely, parathyroid adenomas, gastroenteropancreatic neuroendocrine tumors, and pituitary tumors.21,22 Patients with a diagnosis of secondary hyperparathyroidism remained excluded from the present analysis. The present study was approved by the Ohio State University Institutional Review Board and all patients provided written informed consent.
All resections were performed by 1 of 2 experienced endocrine surgeons (L.A.S. and J.E.P.). Intraoperative parathyroid hormone monitoring was not performed for patients with a diagnosis of MEN1, because all patients underwent planned bilateral neck exploration and 4-gland exploration unless they had recurrent primary hyperparathyroidism in the setting of prior parathyroidectomy. For patients without MEN1, routine intraoperative parathyroid hormone monitoring was used; if preoperative imaging had localized a suspected parathyroid adenoma, a targeted, unilateral parathyroidectomy was undertaken. A decrease in intraoperative serum parathyroid hormone level by more than 50% to a level within the reference range was used as the threshold to conclude further exploration. All tissues deemed to have significant autofluorescence were resected as potential parathyroid tissue and sent for pathologic analysis. Intraoperative frozen section pathologic analysis of resected presumed parathyroid adenoma specimens to confirm the presence of hypercellular parathyroid tissue was performed routinely for patients with MEN1 and selectively, at the operating surgeon’s discretion, for those without MEN1.
Intraoperative imaging was conducted with the handheld NIR PDE-NEO II camera after mobilization of the thyroid gland, with the camera placed 5 cm above the surgical field. Details of the camera NIR specifications and image acquisition procedure have been described previously.15 Images were acquired with ambient light and at NIR wavelengths both in situ and then ex vivo for all resected specimens.
Post hoc, blinded imaging analysis of all prospectively captured intraoperative NIR images was done using Image J software, version 1.52a (National Institutes of Health). Fluorescence values were quantified from representative areas of 10 × 10 pixel dimensions from the regions of maximum autofluorescence of the parathyroid gland, thyroid gland, and adjacent soft tissue.15 Absolute values of maximum autofluorescence, the difference in parathyroid vs background autofluorescence, and the ratio of parathyroid autofluorescence to background autofluorescence were calculated for each patient’s captured images in duplicate and averaged. Positive autofluorescence was defined as parathyroid autofluorescence to background autofluorescence ratio greater than 1.10. The NIR imaging results were then correlated with final pathologic analysis results of all resected specimens.
The primary end points were the rates of false-negative and false-positive autofluorescence. False-negative autofluorescence rate, that is, lack of significant parathyroid gland autofluorescence compared with the background thyroid and/or soft tissue autofluorescence, was defined as a parathyroid autofluorescence-background autofluorescence ratio less than 1.1. False-positive autofluorescence was defined as aberrant autofluorescence of nonparathyroid tissue, which was resected and confirmed on pathologic analysis to not be parathyroid gland.
Demographic and clinicopathologic variables were prospectively collected and retrospectively analyzed. Statistical analyses were undertaken with SPSS, version 23 software (IBM Inc). Comparisons between MEN1 and non-MEN1 cohorts were reported as effect size and 95% CIs. Given the small sample sizes, the estimated effect size was calculated using bias-corrected Hedges g.
A total of 71 consecutive patients underwent resection for a diagnosis of primary hyperparathyroidism. This cohort included 6 patients with a clinical diagnosis of MEN1 who underwent resection of 13 parathyroid glands, and 65 patients without MEN1 who underwent resection of 75 parathyroid glands in total. The 6 patients with a diagnosis of MEN1 included 2 patients who had recurrent hyperparathyroidism due to recurrent growth of hyperplastic parathyroid tissue following previous subtotal 3.5 gland parathyroidectomy; the remaining 4 patients with MEN1 underwent routine subtotal 3- or 3.5-gland parathyroidectomy. Two of the 6 patients with MEN1 had confirmed genetic mutations in the MEN1 gene; all 6 patients met clinical diagnostic criteria for MEN1 syndrome.
Clinicopathologic variables are summarized in Table 1. Most patients in both cohorts were women (MEN1: 4 [67%]; non-MEN1: 51 [78%]). The median (interquartile range [IQR]) age of 49.0 (38.0-53.8) years for patients within the MEN1 cohort was significantly younger than the median (IQR) of 61.0 (54.0-67.0) years within the non-MEN1 cohort (g = −1.04; 95% CI, −1.89 to −0.19). Serum preoperative calcium levels (g = −0.59; 95% CI, −1.43 to 0.25) and resected parathyroid gland size (g = −0.05; 95% CI, −0.55 to 0.64) and weight (g = 0.15; 95% CI, −0.46 to 0.76) on final pathologic examination were similar between the 2 cohorts. The median preoperative serum parathyroid hormone level was higher among patients with MEN1 (171.0 pg/mL; IQR, 117.7-1133.8 pg/mL; conversion to nanograms per liter is 1:1) compared with those without MEN1 (121.9 pg/mL; IQR, 97.6-167.8 pg/mL) (g = 1.45; 95% CI, 0.58-2.31).
Results of post hoc imaging analysis are summarized in Table 2. Median parathyroid gland autofluorescence in situ was significantly lower for patients with MEN1 vs patients without MEN1 (54.4 vs 74.3; g = −1.03; 95% CI, −1.89 to −0.17). Median parathyroid autofluorescence ex vivo was also significantly lower among patients with MEN1 (64.1 vs 86.6; g = –0.80; 95% CI, –1.42 to –0.18) (eFigure 1 and eFigure 2 in the Supplement). The median ratio of parathyroid to background autofluorescence was significantly lower among patients with MEN1 (1.08 vs 1.59; g = –1.59; 95% CI, –2.23 to −0.96), as was the median absolute value of parathyroid autofluorescence – background autofluorescence (6.4 for MEN1 vs 30.6 for non-MEN1) (Figure 1) (g = –1.59; 95% CI, –2.25 to –0.94).
The false-negative autofluorescence rate, ie, the rate of nonfluorescent parathyroid adenomas, was 50% among the MEN1 cohort (3 of 6 patients) (Figure 2A and B) vs 9% among the non-MEN1 cohort (of 65 patients) (Figure 1). The false-positive autofluorescence rate, ie, aberrant increased autofluorescence of nonparathyroid tissue confirmed on final pathologic testing, was 83% among the MEN1 cohort (5 of 6 patients) vs 5% among the non-MEN1 cohort (3 of 65 patients). In all cases of false-positive autofluorescence, both intraoperative frozen section analysis and final pathologic confirmatory analysis of these resected specimens were consisted with benign fibroadipose tissue with no evidence of parathyroid glandular tissue (Figure 2C and D). For the MEN1 cohort, the estimated sensitivity of parathyroid detection by autofluorescence was 25.0% (95% CI, 0.6%-80.6%) and the estimated specificity was 37.5% (95% CI, 8.5%-75.5%). The estimated sensitivity and specificity for the non-MEN1 cohort, by comparison, were 91.2% (95% CI, 81.8%-96.7%) and 95.2% (85.5%-99.0%).
A subset analysis comparing NIR imaging results of patients with MEN1 (n = 6) with those of patients without MEN1 with 4-gland hyperplasia confirmed on pathologic testing (n = 8) demonstrated results similar to the overall comparison of patients with MEN1 vs all those without MEN1 (Table 3). Median parathyroid gland autofluorescence was lower among patients with MEN1 vs patients with 4-gland hyperplasia without MEN1, both in situ (54.4 vs 74.5; g = −1.21; 95% CI, −2.28 to −0.13) and ex vivo (64.1 vs 87.5; g = −1.16; 95% CI, −1.97 to −0.35). Within this subset analysis, the median ratio of parathyroid to background autofluorescence similarly remained lower among patients with MEN1 (1.08 vs 1.76; g = −1.88; 95% CI, −2.76 to −1.00), as did the median absolute value of parathyroid autofluorescence – background autofluorescence (6.4 vs 31.4; g = −1.59; 95% CI, −2.45 to −0.74).
The present study evaluated the results of real-time intraoperative NIR autofluorescence imaging in patients with MEN1 undergoing surgery for primary hyperparathyroidism. Parathyroid adenomas among patients with MEN1 exhibited significantly lower autofluorescence in situ and ex vivo compared with patients with primary hyperparathyroidism without MEN1. The ratio of parathyroid autofluorescence to background autofluorescence was significantly lower for patients with vs without MEN1, leading to high false-negative rates (50%) of parathyroid autofluorescence among patients with MEN1. In addition, significant aberrant autofluorescence of nonparathyroid tissue was associated with high rates of false-positive autofluorescence (83%) among patients with MEN1.
To our knowledge, this study represents the first analysis of parathyroid autofluorescence in patients undergoing surgery for primary hyperparathyroidism in the setting of MEN1 syndrome. Since the initial discovery of intrinsic autofluorescence of the parathyroid gland at NIR wavelengths, significant research has been undertaken to develop handheld imaging technology that could use this phenomenon to improve intraoperative, real-time parathyroid identification and localization without the need for contrast agent injection.19 Early studies focusing on the efficacy of NIR autofluorescence detection during thyroidectomy to minimize unnecessary removal of normal parathyroid glands reported improvements in rates of parathyroid identification and decreases in rates of parathyroid autotransplantation and postoperative hypocalcemia.6,9-11,13 Additional analyses including patients undergoing parathyroidectomy similarly demonstrated improved rates of parathyroid localization with the addition of NIR autofluorescence detection.1,3,5,12 Preliminary results of an ongoing prospective clinical trial demonstrated the efficacy of parathyroid gland identification by NIR autofluorescence.15 Recent US Food and Drug Administration approval in late 2018 of 2 devices for real-time localization of parathyroid glands using NIR autofluorescence detection was based on these encouraging results and may lead to wider adoption of this intraoperative imaging technology.
Although previous reports have suggested that both normocellular parathyroid glands and hypercellular adenomatous glands exhibit intrinsic NIR autofluorescence, to our knowledge, patients with MEN1 have not been analyzed within these study cohorts.14,15 Limited analysis of patients with secondary hyperparathyroidism associated with chronic kidney disease suggested that these patients may exhibit lower intrinsic parathyroid autofluorescence than patients with primary hyperparathyroidism, perhaps as a result of differential expression of unknown receptors or proteins within parathyroid tissue.5 Although patients with secondary hyperparathyroidism were excluded from the present analysis, the diffuse, multiglandular parathyroid hyperplasia seen among patients with MEN1 may be similar to the pathophysiology and histologic changes observed in parathyroid glands of secondary hyperparathyroidism. Hyperparathyroidism is the most common clinical manifestation of MEN1, occurring in 95% of patients.23 Thus, nearly all patients with a diagnosis of MEN1 will require surgical intervention for primary hyperparathyroidism, making the applicability of intraoperative NIR autofluorescence technology to this patient population a clinically relevant issue.
Anecdotal observations suggested that autofluorescence of parathyroid glands might be less conspicuous in patients with MEN1. We sought to quantitatively analyze autofluorescence values for patients with and without MEN1 with primary hyperparathyroidism and correlate these imaging findings with histopathologic test findings. The results of the present preliminary analysis demonstrated significantly lower quantified absolute values of parathyroid autofluorescence in situ and ex vivo and significantly lower parathyroid-to-background autofluorescence ratios for patients with vs without MEN1. On a subset analysis of patients with similar glandular pathophysiologic characteristics, namely histopathologic testing-confirmed 4-gland hyperplasia, patients with MEN1 again demonstrated lower parathyroid autofluorescence and lower parathyroid-to-background autofluorescence ratio compared with patients without MEN1. These quantifiable findings translated to a 50% false-negative autofluorescence rate for patients with MEN1; in other words, 3 of 6 patients had false-negative NIR imaging of histopathologic-confirmed hypercellular parathyroid glands. The frequent finding of pronounced, quantifiably positive autofluorescence of nonparathyroid tissue among patients with MEN1 was unexpected. Five of 6 patients with MEN1 were found to have autofluorescence of tissue thought to be potential parathyroid gland that was then resected and, on pathologic analysis, found to be benign, nonparathyroid fibroadipose tissue. This finding translated to an 83% false-positive autofluorescence rate among patients with MEN1 vs only 5% among patients without MEN1. One possible explanation may be the difference in age of the patients, as patients with MEN1 were significantly younger individuals who may have a greater amount of brown fat. The high rate of false-positives may also be associated with the more extensive surgery that the patients with MEN1 underwent because they often had a 4-gland exploration with removal of additional central neck tissue and thymus to capture any ectopic parathyroid tissue. A strength of the present study was that, unlike most prior studies, all patients had pathologic confirmation of all specimens to correlate with intraoperative imaging findings.
The study has limitations. An inherent limitation of the present study is that all captured images underwent post hoc quantification using imaging software. The existing technology for this camera system does not have the capabilities for real-time, intraoperative quantification of NIR images. The results of this preliminary study are also limited by the small size of the single-institutional cohort and the relative rarity of MEN1; plans are underway to attempt to analyze the utility of autofluorescent real-time NIR imaging in a larger, multi-institutional format. In addition, the 95% CIs for many of the effect sizes were fairly wide, and whether these results would be consistent with larger study cohorts remains to be seen. Nevertheless, the visual differences in parathyroid gland autofluorescence between the MEN1 and non-MEN1 cohorts are striking and the quantified differences in autofluorescence variables between the 2 cohorts appear to be clinically significant and worthy of further investigation.
Lower intrinsic parathyroid autofluorescence, as well as increased aberrant autofluorescence of nonparathyroid tissue, was observed among patients with MEN1 undergoing resection for primary hyperparathyroidism in this preliminary report. This finding resulted in high rates of false-negative autofluorescence and false-positive autofluorescence among the cohort of patients with MEN1. Intraoperative identification of parathyroid glands using their intrinsic autofluorescence by real-time intraoperative NIR imaging holds significant promise but may have limited utility among the specific subset of patients with primary hyperparathyroidism and MEN1. Further investigation of NIR imaging characteristics within this patient population is warranted.
Accepted for Publication: June 9, 2019.
Corresponding Author: John E. Phay, MD, Division of Surgical Oncology, Department of Surgery, The Ohio State University Wexner Medical Center, 410 W 10th Ave, N-907 Doan Hall, Columbus, OH 43210 (email@example.com).
Published Online: August 1, 2019. doi:10.1001/jamaoto.2019.1987
Author Contributions: Drs Squires and Phay had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Squires, Shirley, Phay.
Acquisition, analysis, or interpretation of data: Squires, Shen, Jarvis, Phay.
Drafting of the manuscript: Squires, Shirley, Phay.
Critical revision of the manuscript for important intellectual content: Squires, Shen, Jarvis, Phay.
Statistical analysis: Squires, Shen.
Administrative, technical, or material support: Jarvis, Phay.
Supervision: Shirley, Phay.
Conflict of Interest Disclosures: Dr Phay reported a licensing agreement from AI BIomed outside the scope of the submitted work and had a patent to Intra-operative Use of Fluorescence Spectroscopy and Applications of Same licensed. No other disclosures were reported.
Meeting Presentation: The study was an oral presentation at the 2019 Academic Surgical Congress; Houston, Texas; February 6, 2019.