Correlation of subtraction parathyroid scintigraphy with weight , pathohistologic finding and oxyphil cell content of parathyroid glands in parathyroid hyperplasia

Background/Aim. Parathyroid hyperplasia (PHP) is defined as an absolute increase in the mass of parenchymal cells of the parathyroid gland. PHP is classified as primary, secondary and tertiary. The enlargement of parathyroid glands (PG) is usually asymmetric, resulting in a “dominant“ gland. In order to confirm the diagnosis, at least two glands should be examined histologically. Subtotal parathyroidectomy, i.e. removal of the three PG and leaving a small remnant of the forth, is the treatment of choice. High percent of PHP recurrence imposes the need for preoperative high sensitivity localizing procedures. Parathyroid scintigraphy localizes about 60% of hyperplastic glands. The aim of this study was to correlate findings of subtraction parathyroid scintigraphy (SPS) with weight, pathohistologic finding and oxyphil cell (OC) content of PG in patients with primary, secondary and tertiary parathyroid hyperplasia. Methods. Twenty-seven patients with primary/secondary PHP underwent SPS before surgery. Scintigraphic results were graded from 1–5, in relation to the degree of uptake. SPS graded 3, 4 and 5 were considered positive. The number and weight of operated PG were evaluated macroscopically. Pathohistologic and cellular types were defined on standard stained hematoxylin-eosin slides. OC content was defined as a percent of OC and graded from 1 to 3: grade 1 < 10%, grade 2 10% and grade 3 20% of OC. Results. SPS localized dominant gland in all patients with sensitivity 100%, and 51 from 73 hyperplastic PG, with sensitivity per gland of 70%. PG weighed 0.1 g to 6.7 g (median 1 g). A significant positive correlation (p < 0.0001) was found between the SPS results and PG weight. A significant positive correlation was found between PG weight and OC content (p = 0.0002). An insignificant correlation was found between SPS and OC content. Thirty-eight PG had < 10% of OC, 32 PG had 10% and 3 PG had 20% of OC. Four patients had diffuse PHP and 23 patients nodular PHP. There was no statistically significant difference in SPS results compared to hyperplasia type, and between OC content and hyperplasia type. A significant positive correlation (p = 0.05) was found between PG weight and hyperplasia type. Conclusion. A high positive correlation was found between SPS results and PG weight, PG weight and OC content and PG weight and hyperplasia type. Between SPS results and OC content, and between SPS results and hyperplasia type, an insignificant correlation was found. Our results showed that SPS is a reliable and very sensitive diagnostic tool in detecting abnormal PG in parathyroid hyperplasia, reaching 100% sensitivity in detecting a “dominant gland” and sensitivity per localized gland of 70%. Causes that affect increased uptake of liposolubile Tc99m radiopharmaceuticals (RF) in the hyperfunctional PG tissue and conditions which prevent RF admission into the PG cells still remain to be accurately and precisely determined.

1 g).A significant positive correlation (p < 0.0001) was found between the SPS results and PG weight.A significant positive correlation was found between PG weight and OC content (p = 0.0002).An insignificant correlation was found between SPS and OC content.Thirty-eight PG had < 10% of OC, 32 PG had 10% and 3 PG had 20% of OC.Four patients had diffuse PHP and 23 patients nodular PHP.There was no statistically significant difference in SPS results compared to hyperplasia type, and between OC content and hyperplasia type.A significant positive correlation (p = 0.05) was found between PG weight and hyperplasia type.Conclusion.A high positive correlation was found between SPS results and PG weight, PG weight and OC content and PG weight and hyperplasia type.Between SPS results and OC content, and between SPS results and hyperplasia type, an insignificant correlation was found.Our results showed that SPS is a reliable and very sensitive diagnostic tool in detecting abnormal PG in parathyroid hyperplasia, reaching 100% sensitivity in detecting a "dominant gland" and sensitivity per localized gland of 70%.Causes that affect increased uptake of liposolubile Tc99m radiopharmaceuticals (RF) in the hyperfunctional PG tissue and conditions which prevent RF admission into the PG cells still remain to be accurately and precisely determined.

Introduction
Parathyroid hyperplasia (PHP) is defined as an absolute increase in mass of parenchymal cells of the parathyroid gland (PG).It is found in 7% of the PGs examined during a routine autopsy.At least two glands should be examined histologically in order to establish the diagnosis.Macroscopically PG can be of normal size or significantly enlarged.All glands can be affected equally, but enlargement is usually asymmetric, resulting in a "dominant" gland 1 .Parathyroid hyperplasia combines hyperplasia of the chief cells and oxyphil cells (OC) and it most commonly occurs secondary to hyperplasia of the chief cells.Hyperplasia can be diffuse and nodular.Parathyroid hyperplasia is classified as either primary (pPHP), secondary (sPHP) or tertiary (tPHP) correspondingly to primary (pHPT), secondary (sHPT) and tertiary hyperparathyroidism (tHPT) [2][3][4][5][6][7][8][9][10][11] .
Primary parathyroid hyperplasia occurs in approximately 15% of patients with pHPT.Most cases are sporadic.Approximately 20% of pPHP is the result of chief-cell hyperplasia and is associated with some of hereditary syndromes, most importantly multiple endocrine neoplasia syndromes one and two (MEN1 and MEN2).PGs were altered in 90% of MEN1 and 30-40% of MEN2 cases 7,9,[12][13][14][15] .
Secondary PHP occurs within sHPT, most commonly as a result of renal insufficiency.Renal failure influences phosphate excretion resulting in decrease of calcium level and increased parathyroid hormone (PTH) secretion.Secondary PHP can also occur due to: vitamin D deficiency, hypomagnesaemia, malnutrition, urinary calcium excretion, influence of certain medicaments.Tertiary PHP represents an autonomous PG function in patients with previously established sHPT.The existence of tHPT is typical for patients with chronic renal failure on dialysis 6,16 .
Surgical removal of the PG is the most successful treatment of HPT.Classic procedure involving bilateral exploration and dissection of the offending PG in the neck has a 95% success rate in all patients.Failure to localize the ectopic PG and undiagnosed multiple PG disorder in pHPT represents the main reason for insufficiently successful surgical intervention.Parathyroid surgery was also indicated for patients with hipercalcemia, high PTH level and/or renal osteodystrophy in sPHP which cannot be successfully medicated.Another criterion for surgery is the detection of enlarged PG by medical imaging.Subtotal parathyroidectomy, i.e. removal of the three PG and leaving a small remnant of the fourth is the treatment of choice, even though removal of all PG with autotransplantation of the parathyroid tissue in the sternocleidomastoid muscle or the forearm is also recommended 17,18 .
Introduction of ultrasound diagnostics and subtraction parathyroid scintigraphy (SPS) via liposolubile radiopharmaceuticals (RF) i.e.Tc99m-MIBI, and Tc99m-tetrofosmin (TRF) has significantly increased the detection of the enlarged PG in the last ten years.Diagnostics facilitates surgical procedures and also decreases the recurrence and the need for reoperation 17 .The main advantage of subtraction parathyroid scintigraphy (SPS) is a high sensitivity (up to 97%) and specificity (up to 95%) in localizing solitary pHPT adenoma 19,20 .Scintigraphy sensitivity varies in pPHP, sPHP and tPHP, rating between 15% and 60% 21 .Small adenomatous and hiperplastic PG can incite false negative findings.False positive findings are usually caused by coincidental nodus of the thyroid gland.A large number of analyses indicates a positive correlation between scintigraphic detection and the size of hyperfunctional PG.A number of analyses suggests that the intensity of Tc99m-MIBI uptake is in positive correlation with nodular hyperplasia and OC content in PG.A correlation of positive scintigraphic findings with weight and pathohistologic finding of the hyperfunctional PG has been the subject of a number of studies [22][23][24][25][26][27][28] .Inability to recognize hyperplasia in pHPT and a high percentage (10-70%) of disease recurrence within sHPT and tHPT requires determination of factors which influence the increase in RF uptake intensity and PG scintigram sensitivity, particularly at hyperplastically altered glands.
The aim of this study was to correlate findings of subtraction parathyroid scintigraphy (SPS) with weight, pathohistologic finding and oxyphil cell (OC) content of the PG in patients with primary, secondary and tertiary parathyroid hyperplasia.

Methods
The research included 27 patients operated for pHPT or sHPT at an average age of 50 (27-74) years with a total number of 73 hyperplastically altered PG.All of the patients underwent SPS before surgery.Seven patients at an average age of 64.4 (52-74) years had pHPT and 20 patients at an average age of 44.7 (20-73 years) had sHPT.Eighteen patients underwent SPS with Tc99m-MIBI, while 9 patients underwent SPS with Tc99m-tetrofosmin.Static scintigrams of the head, neck and chest were performed 15 min after iv injection of 740 MBq of Tc99m-MIBI/TRF.Anterior projection images were obtained using an ADAC gamma camera with a low-energy, high-resolution collimator in zoom mode in 128 × 128 matrix size with a 20% energy window, connected.2 000 000 impulses per position were gathered.Later scintigrams of the head, neck and chest were performed 2 h and 3 h after iv injection of Tc99m-MIBI/TRF.Four to 24 h iv injection of Tc99m-MIBI/TRF after washout of RF from the PG and the thyroid gland, iv injection of 185 MBq Tc99m-pertechnetat (deposited exclusively in thyroid gland) was administered.Ten minutes after iv injection of Tc99m-pertechnetat, static scintigrams of head and neck in the anterior projection were performed, gathering 2 000 000 impulses per position.After normalization of early (Tc99m-MIBI/TRF) scintigram and (Tc99m) scintigram and motion correction, subtraction Tc99-petrechnetat from Tc99m-MIBI/Tc99m-tetrofosmin scintigram was performed.Zones of the increased uptake of Tc99m-MIBI/TRF visible in subtraction image represent the hyperfunctional tissue of the enlarged and altered PG.Scintigraphic (SPS) findings were graded 1-5: grade 1 -normal RF uptake, grade 2 -discretely increased RF uptake, grade 3 -slightly increased RF uptake, SPS grade 4 -increased RF uptake, 5 -clearly increased RF uptake.SPS grades 1 and 2 were considered negative findings ( ).SPS results grade 3 were evaluated as suspicious findings (±), while grades 4 and 5 were considered positive findings (+).
Number, weight and mass of the removed PGs were analyzed macroscopically, while intraoperative, ex tempore analysis was done microscopically on standard hematoxylineosin stained slides in order to determine the substrate: adenoma, type of hyperplasia, distribution and substitution of the main, light, OC and degenerative alterations.OC percentage was graded: 1 -for < 10%, grade 2 -for 10%, and grade 3 -for 20% of OC by a standard preparation method.
SPS sensitivity was calculated by the equation used to calculate matrix sensitivity.A correlation between SPS findings and PG weight; SPS and OC percentage; PG weight and OC percentage was determined using the Spearman's rank correlation test.Differences in the pathohistologic findings (diffuse or nodular hyperplasia) in relation to: scintigraphic finding, PG weight and OC number, were determined using the rank sum test (Mann-Whitney).A significance level was p < 0.05.

Results
Positive SPS results were recorded in all of the patients, i.e. at least a single enlarged PG was localized (scintigraphy sensitivity per patient was 100%).SPS localized 48 out of 73 hyperplastic PG (SPS grade 5-39 PG, SPS grade 4-9 PG), 22 glands were not visible (SPS grade 2-4 PG, SPS grade 1-18 PG), while findings in 3 PG were evaluated as suspicious, SPS grade 3. SPS sensitivity per localized gland, along with 3 suspicious findings, was 70% (Figure 1-3).The PG in patients with positive and suspicious scintigraphy results weighed from 0.1 g to 6.7 g (median 1g).There was no statistically significant difference in the weight of the hyperplastic PG in relation to the type of hyperparathyroidism.PG weight in patients with pHPT went from 0.2 g to 2 g (median 0.75 g), in patients with sHPT it went from 0.1 g to 6.7 g (median 0.8 g).
A significant positive correlation (p < 0.0001) between SPS results and weight of the hyperplastic PG was found.Glands invisible in scintigraphic scanning weighted between 0.1 g and 1 g (median 0.5 g).In patients with positive and suspicious SPS results, PG weighted from 0.1 g to 6.7 g (median 1 g).Glands invisible in scintigraphic scanning (SPS grades 1 and 2) were significantly smaller than PG localized by scintigraphy (SPS grades 3, 4 and 5) (p < 0.001) (Figure 4).
A significant positive correlation between PG weight and OC percentage (p < 0.001) was also noted.Thirty-eight PG had < 10% of OC, thirty-two PG had 10% of OC and three PG had 20% of OC in them.A notable percentage of OC was found in hyperplastically altered PG of larger weight (Figure 5).
There was no significant correlation between SPS results and percentage of OC in PG.Positive SPS results noted 23 PG with < 10% of OC, 24 PG 10% of OC and 3 PG 20% of OC.False negative SPS results (significantly smaller PG) noted 15 PG with < 10% of OC, 8 PG 10% of OC.Even though there was no statistically significant positive correlation found between SPS results in relation to OC percentage, 65% of small weight PG with SPS grades 1 and 2 (negative findings) had < 10% of OC.Four patients had diffuse PHP (2 patients with pHPT and 2 with sHPT) and 23 patients nodular HPH (5 patients with pHPT and 18 with sHPT).There was no statistically significant difference in SPS results in relation to hyperplasia type.
There was no significant difference between OC percentage and hyperplasia type.
A significant difference in PG weight and hyperplasia type (p < 0.05) was determined.The weight of the PG with diffuse hyperplasia was significantly smaller than the weight of PG with nodular hyperplasia.

Discussion
Tc99m-MIBI and Tc-99m-tetrofosmin were primarily introduced as RFs used for myocardial perfusion scintigraphy.Uptake of the two RFs is recorded in certain malignant tumors, in the PG as well as in the functional thyroidal tissue.Tc99m-MIBI and Tc-99m-tetrofosmin are intracellular markers 17,19,20,[28][29][30][31] .Their liposolubility enables them to travel through cellular membrane and to enter the cell.Mitochondria are responsible for RF cellular uptake, but a complete mechanism of their binding and release (perfusion, metabolic activity and phases of the cell cycle) are still relatively unknown 24,29 .Elimination of RFs from the normal thyroid tissue is usually much faster than elimination from the hyperfunctional PG which allows PG visualisation on delayed, i.e. late, scintigrams, one or more hours after iv injection of RFs.Subtraction scintigraphy allows visualisation of glands with high RF uptake and high release rate, visualisation of intrathyroidal PG, and it also allows us to differentiate thyroid gland nodules from the enlarged PG.The hyperplastic PG can be of normal size or significantly enlarged.All of the glands can be equally affected, but the enlargement is most commonly asymmetric, resulting in a "dominant" gland.Smaller size of hyperplastically altered glands in relation to adenomatously altered PG is the main reason for the lower sensitivity of scintigraphic detection of hyperplasia 21,22,25,27,28,32 .In our study, SPS sensitivity in localization of the largest gland was 100%, while the sensitivity in determining all of the hyperplastically altered glands was 70%.Palestro et al. 32 assume that the lower sensitivity in sPHP was caused by smaller amount of OC and the number of mitochondria which results in lower uptake and faster washout of RF from the cells.A number of analyses resulted in positive correlation between SPS and size, i.e. weight, of the hyperfunctional PG, regardless of the type of disease (adenoma, hyperplasia or carcinoma) 21,22,[24][25][26][27][28][32][33][34] . In our atients, we established a positive correlation between SPS results and size of the hyperplastically altered glands, both in pPHP and sPHP.
A large number of analyses determined that the intensity of Tc99m-MIBI uptake is in positive correlation with nodular hyperplasia 24,[26][27][28]32 . Muro et al.21 proved no correlation between secondary hyperplasia and SPS results.Correspondingly, there was no significant positive correlation between type of hyperplasia and positive SPS results in our patients. Th results could be partly explained by a small number of PG with diffuse hyperplasia (13 patients, 6 with pHPT and 7 with sHPT), as well as by a significantly smaller PG weight.
In late 1960s, Christie 36 stated that OC of the normal PG have no secretory function and that they represent the chief cells in gland involution.A number of papers proved active secretory function of OC, both in normal and hyperfuntcional PG 4,33,35,36 .OC cytoplasm is composed of tightly packed mitochondria and glycogen 16,36 .A number of authors confirmed a positive correlation of SPS with OC content in PG 22, 24, 25, 32- 34 .Yamaguchi et al. 37 , as well as Bhatnagar et al 28 ; Piñero et al. 38 and Ugur et al. 39 found no positive correlation between the number of OC and positive SPS results.Our group of patients also proved no positive correlation between positive SPS results and OC content in hyperplastic PG.
Apart from the factors which influence uptake kinetics of liposoluble Tc99m RF in hyperplastically altered PG, a number of other factors that might influence kinetics are being tested.This is primarily applied to factors which prevent admission and uptake of RF in PG cells.Individual histological characteristic and metabolic activity of the hyperfuntctional parathyroid tissue influence the positivity of Tc99m-MIBI and Tc99m-tetrofosmin SPS results 9,26,34,38 .Furthermore, unrecognized hyperfuntcional PG were conditioned by predominance of light chief cells 22,[24][25][26]40 . The btained results are still questionable.
RF kinetics can be altered by different serum calcium levels due to its influence on the membrane potential 23,34,41 and mitochondrial content in PG cells 41 .It has been proven that cellular expression of P-glycoprotein (P-gp) and multiple drug resistance protein (MDR) can be a significant factor in obtaining false negative SPS results in HPT 28,37,42 .Concentration of MIBI and tetrofosmin in altered PG cell is in negative correlation with P-gp 28,37,42 .Positive correlation between the intensity of Tc99m-MIBI uptake and phases of cell cycle will be the subject of future analyses.High level of RF uptake as indicator of metabolic status is in positive correlation with the phases of positive growth of hyperfunctional parathyroid tissue 43 .Complete mechanisms of uptake and washout of Tc99m-liposolubile RF in PG are still missing and are the subject of many studies.

Conclusion
In this study the PG with diffuse hyperplasia was of significantly smaller weight than the PG with nodular hyperplasia.A significant positive correlation between SPS grade and the weight of hyperplastic PG was found.A positive correlation was also established between PG weight and OC content, but there was no statistically significant difference between SPS results depending on OC content and type of hyperplasia.Our results showed that SPS is a reliable and very sensitive diagnostic tool in detecting the abnormal PG in parathyroid hyperplasia, reaching 100% sensitivity in detecting a "dominant gland" and overall sensitivity of 70%.Causes that affect increased uptake of liposolubile Tc99m radiopharmaceuticals in the hyperfunctional PG tissue and conditions which prevent RF admission into the PG cells still remain to be accurately and precisely determined.

Fig. 4 -
Fig. 4 -Graphical display of scintigraphic assessment findings in relation to the weight of the operated parathiroid glands (PG); x-axis: number of PG with the appropriate grade of scintigraphic findings (subtraction parathyroid scintigraphy -SPS), y-axis: weight of PG in grams.

Fig. 5 -
Fig. 5 -Graphical display of parathiroid glands (PG) weight in relation to the percent of oxyphil cells (% OC) of the operated parathiroid glands; x-axis: number of PG with the appropriate grade of % OC, y-axis: weight of PG in grams.