Evaluation of factors predicting postoperative recurrence and metastasis of parathyroid cancer: a single-center study

Parathyroid carcinoma (PC) is a rare endocrine tumor accounting for 0.5–6% or less in cases of primary hyperparathyroidism [1‐3]. Currently, the World Health Organization (WHO) histopathologic findings are the standard methods for diagnosing the malignancy of parathyroid neoplasms. According to the above, the histopathologic criteria for PC include the presence of vascular invasion with capsular invasion and progression and/or metastasis to adjacent tissues [4]. However, the small number of cases of PC makes it difficult to predict prognosis in the postoperative period of PC.

The 5- and 10-year overall survival (OS) rates for PC patients are estimated to be 78–85% and 49–70%, respectively [5‐7]. Parathyroid cancer is an endocrine tumor with poor prognosis presenting severe clinical findings such as recurrence and metastasis. Histopathologic criteria of parathyroid cancer proposed by the World Health Organization (WHO) alone are insufficient to predict the degree of malignancy, such as of postoperative recurrence and distal metastasis. Recently, a nationwide survey examined prognostic factors in parathyroid carcinoma; Kong et al. reported that three factors were prognostic as follows: age > 50 years, parathyroidectomy only, and previous reoperation [8]. Other reports indicate that tumor size, distant metastasis, and other factors define prognosis [9]. However, because of the small number of cases of parathyroid cancer at a single institution and the paucity of reports, even in multi-institutional studies, there is a need to examine factors that predict prognosis in the postoperative period of parathyroid cancer.

Characteristics of PC include severe hypercalcemia, palpable mass in the cervical region, intraoperative findings of localized invasion, and suspicion of distant metastasis by preoperative examination [3, 10, 11]. Aspiration biopsy cytology of the parathyroid gland for diagnosis carries the risk of causing cell dissemination, and therefore, as a rule, it is not recommended [12, 13]. In general, parathyroid carcinoma and parathyroid adenoma are difficult to distinguish before surgery because the clinical presentation, biochemical findings, and imaging findings are almost identical to each other; hence, most patients are presumed to have a benign parathyroid adenoma, and a surgery is performed on them. It is also difficult to distinguish the parathyroid carcinoma from the parathyroid adenoma at the time of surgical resection. Consequently, it is not uncommon for cancer to be diagnosed by histopathologic examination after surgery [1, 14]. Most patients who develop localized recurrence and distant metastasis die due to uncontrollable severe hypercalcemia [5, 15]. Therefore, curative treatment involves performing total resection by surgery and reliably removing the lesion, making it undoubtedly the only desired treatment [3, 15‐17]. When postoperative recurrence occurs, surgical resection has been reported to have prolonged survival [18]. Reoperation at recurrence is therefore an effective approach.

Following surgery, once PC has been diagnosed by histopathologic diagnosis, the goal in the management of PC is to detect recurrence of the disease early and to reduce death by aggressive treatment. In actual clinical practice, some cases are encountered that have good progress with no recurrence and no distant metastasis. Therefore, predictors of recurrence and/or distant metastasis are needed to distinguish between cases that follow a slow course and those that should be followed closely. To distinguish cases of such gradual progression and cases with severe progression that should be observed, accurate risk stratification of PC is needed. We hypothesized that by searching for prognostic predictors of recurrence and/or distant metastasis based on clinicopathologic examination results, we could identify cases that require aggressive postoperative management. Therefore, we aimed to examine factors that serve as indicators to determine prognosis following surgery for PC.

Determining the prognosis of PC accurately and effectively is clinically significant for subsequent follow-up. However, owing to the rarity of PC, there is little clinical evidence regarding prognosis, and because there is no approved staging system, clinicians have difficulty determining prognosis accurately. However, we experienced cases that presented gradual clinical findings only and that progressed postoperatively without recurrence and metastasis. According to research in recent years using a population-based database, several reports have investigated prognostic factors of PC patients. Sadler et al. analyzed the National Cancer Data Base (NCDB) and found that lymph node metastasis positivity and age are associated with reduced survival rate [16]. Like Sadler et al., Asare et al. analyzed the NCDB and reported that patient age, tumor size, and sex affect the survival of PC patients [20]. Although the two reports used the same database, we believe that a difference arose in the analysis results because the subject samples included different age groups.

Definitive diagnosis of PC can only be made by histological examination following surgical resection. Schantz and Castleman first reported the pathological criteria defining PC in 1973 [21], the characteristics of which included capsular invasion, vascular invasion, fibrous bands, and mitotic activity. Thereafter, according to the WHO classification that was revised in 2017, the diagnosis of malignant tumors should be limited to the following: (1) tumors showing invasive proliferation into the thyroid gland and soft tissue with confirmed capsular and/or extracapsular vascular invasion and metastasis, (2) vascular invasion occurring when the tumor invades the capillary vessels or surrounding soft tissue vessels, and (3) tumor cell properties are often similar, with cell populations formed by broad bands of fibrous connective tissue extending from the capsule surrounding the tumor. Zonal histology is present in 90% of PCs and mitotic images in 80%; however, both characteristics are not considered to be specific to malignant tumors [19]. In immunohistological staining, Ki-67 of 5% or above suggests PC [22‐24]. In terms of prognosis, Kameyama et al. reported elevated Ki-67 levels in PC patients with a poor prognosis [25], and Iihara et al. suggested that Ki-67 of 5% or above indicates a poor prognosis, and in particular, that recurrence develops early after surgery when Ki-67 is 10% or above [26]. Moreover, Lenshow et al. demonstrated that a Ki-67 index of 10% increases the postoperative recurrence rate% [27]. In our study also, we confirmed that the Ki-67 index is a prognostic factor for recurrence and metastasis. Patients with a Ki-67 index of 4.1% or higher had a higher incidence of recurrence and distant metastasis than those with a Ki-67 index of less than 4.1% (p < 0.001). With regard to the Ki-67 index, though our results were lower than the cutoff value previously reported, we believe that this may be because we used the occurrence of recurrence or distant metastasis as a criterion in this study in order to examine the indicators for early detection of recurrence and individualized treatment, while the previous report used the survival rate.

Bondeson et al. suggested that findings of coagulation necrosis in pathological findings are related to recurrence of PC [28]. Williams et al. proposed a dataset for PC as a record for future guidelines [29]. They noted that necrosis might be more generalized in tumors of high-grade malignancy and incorporated necrosis as a required element for such guidelines. However, as a point of caution, necrosis should be differentiated from nonspecific infarct-like necrosis, which might be caused by spontaneous infarction, prior episodes of rupture, and preoperative cytodiagnosis. In the present study, patients with necrosis, compared to patients without necrosis, exhibited a higher incidence of recurrence and distant metastasis, thereby demonstrating the usefulness of the findings of necrosis (p < 0.0027).

With regard to the effect of the tumor size on the prognosis of PC, the literature offers different opinions. Asare et al. reported that tumor size is a predictor of distant metastasis, and that the development of distant metastasis affects survival [30]. Lee et al. and Talat et al. reported that tumor size is not an important prognostic marker [3, 6]. On the other hand, Hsu et al. concluded that a tumor size of 30 mm or greater is associated with lymph node metastasis [31]. Furthermore, Asare et al. found a correlation between patients with tumor size of 40 mm or greater and an increased risk of death [20], while Qian et al. reported that among 604 patients with PC in the Surveillance, Epidemiology and End Results (SEER) database, tumor size of 35 mm or greater is associated with worsening of cancer-specific survival (CSS) in patients with PC [9]. In our analysis, there was no significant difference in parathyroid adenoma size, but in parathyroid weight, the results were consistent with the possibility of recurrence or distant metastasis after surgery for PC. Furthermore, in the present study, as a result of having examined the tumor weight measured postoperatively using the cutoff value of 4890 mg from the ROC analysis as a predictor of prognosis (sensitivity 100% and specificity 60.9%), patients with tumor weight ≥ 4890 mg, compared to those with < 4890 mg, had a higher incidence of recurrence and distant metastasis (p = 0.0300). Few reports in the past have examined tumor weight. The reason for this is that when PC is suspected preoperatively, the tumor is resected along with the surrounding organs, especially the thyroid gland, to account for the possibility of disease spread to the surrounding parathyroid tissue. Therefore, as the results of the present study were determined once the parathyroid gland alone was weighed and only serve as a reference, not all cases of PC can necessarily be determined preoperatively, and thus, in cases where tumor weight has been measured, we believe that it is useful as supporting diagnosis of prognosis.

The purpose of the present study was to investigator indicators that can serve to predict prognosis in patients with PC diagnosed postoperatively by histopathologic diagnosis. We examined the incidence of recurrence and distant metastasis according to the number of patients with the risk factors identified in the present study, including Ki-67, pathological findings of necrosis, and tumor weight. As a result, the higher the number of patients with these risk factors, the higher the incidence of recurrence and distant metastasis, and in actual clinical practice, it was possible to identify patients who needed careful follow-up observation. We believe that it will serve as an indicator for early detection and treatment of postoperative recurrence and metastasis of parathyroid cancer and will help improve the prognosis by early treatment.

The present study has several limitations. First, the study is a single-center analysis, which limits the number of cases. Second, uncontrolled or unrecognized bias due to the relatively small sample size is inevitable. Accordingly, to further evaluate the clinical applicability of the present results, a further prospective study is needed by multicenter collaborative research with a larger sample size. However, the fact that a prospective study of extremely rare tumors is not possible, we believe, it is also a limitation of this study. 

The present study demonstrated that there are three factors that predict recurrence and prognosis in patients with PC diagnosed postoperatively by histopathological diagnosis. These three factors were Ki-67 index, weight of the parathyroid gland, and findings of necrosis. Patients with Ki-67 index ≥ 4.1, parathyroid gland weight ≥ 4890 mg, and necrosis findings might develop recurrence and distant metastasis early and require strict follow-up. We believe that the factors above are useful as indicators to predict prognosis following surgery and can help select appropriate treatment.