Prognostic assessment plays a crucial role in the management of TC, and the most commonly used risk assessment system is the one proposed by the ATA in 2015[3]. It uses variables such as primary and metastatic tumor histology from anatomic and pathologic perspectives to classify recurrence risk levels. However, it does not include tumor growth rate, one of the two basic characteristics of tumor cell growth. Over the past 40 years, Ki-67 LI has been extensively evaluated as a prognostic indicator in breast cancer, prostate cancer, and other tumors[6, 7, 14, 15]. Studies have shown that higher Ki-67 LI is related to worse prognosis, the same as in TC[8, 9, 16]. The latest WHO guideline 2022 has pointed out that DHGTC with high-grade pathological features such as increased mitotic index and the presence of tumor necrosis are associated with an aggressive clinical course like tumors classified as PDTC[17, 18]. Usually, immunohistochemical staining of PDTC and DHGTC showed a high Ki67 proliferation index in the range of 10–30%. To avoid the interference of this established fact in our trial, we therefore excluded patients with Ki-67 of > 10% in this subgroup. Clinically and epidemiologically, PDTC and DHGTC are rare, representing less than 1-6.7% of all thyroid cancers[19]. So, we wanted to investigate whether the majority of patients in the thyroid cancer population could be prognosticated using Ki-67 expression (<10%) in order to identify patients with poorer prognosis and provide them with personalized treatment. Interestingly, we found that Ki-67 LI of 5–10% could predict better outcomes after RAI therapy in PTC patients in this study.
Miyauchi et al. showed that higher Ki-67 was significantly related to poorer DFS in PTC patients ,but in their study, only 39.2% patients underwent RAI therapy[8]. However, our current study showed the inverse results, Ki-67 LI of 5–10% was less likely to result in NER after RAI treatment. We found similar reports on oral squamous-cell carcinoma[20], colorectal cancer[21], and rectal/rectosigmoid cancer[22], all demonstrating that high Ki-67 was an indicator of good prognosis. Possible explanations are as follows: (1) Tumor proliferative activity is well known to be heterogeneous[23], and PTC rarely exists as a microscopically homogenous tumor (only the classical variant)[24]. Intratumor heterogeneity depends on the tumor invasion front in opposition to the central tumor compartment. Palmqvist et al. believed lower Ki-67 was associated with poor prognosis, especially at the invasion front[25], but similar research has not been performed in TC. (2) In some cases, tumors with high Ki-67 levels might still exhibit better clinical outcomes because they have intact apoptotic mechanisms. This means that despite their increased proliferative activity, these tumors can still undergo cell death when necessary, which can help limit their aggressiveness and potential for metastasis[9]. Further research is needed to fully understand the mechanisms underlying this discrepancy and its effect on tumor behavior. The slower proliferation rate observed in tumors with high Ki-67 but low mRNA expression might contribute to improved outcomes in colorectal carcinomas. This suggests that factors beyond Ki-67 staining index alone, such as gene expression patterns, help determine the aggressiveness and growth rate[26]. (3) In addition, Siyuan Xu et al. showed that PTC patients with HT showed fewer aggressive tumor characteristics and better prognoses than patients without HT[27]. In this study, we similarly found that patients with high Ki-67 had a higher probability of HT (11/28, 39.9%), which might explain why these patients had good prognoses. Additionally, immunohistochemistry could not measure the true proliferation rate of tumors, since it is a fraction of the growth time required to complete the cell cycle[22].
Some studies have integrated age, sex, tumor size, ETE, LNM, distant metastases, and TNM stages into clinical decision making in RAI treatment for TC[28–32], but few have taken Ki-67 expression into consideration. In this retrospective analysis of 187 PTC patients, we found that Ki-67 LI of 5–10% was related to good outcomes after RAI therapy and RAI treatment might have altered the tumor proliferation process. Our patients with high Ki-67 LI were younger overall (all of them were ≤ 55 years old), younger patients might have a higher survival rate[33]. In addition, other studies’ follow-up time was 146 months[9], 68.5 months[34], or even longer, while our median follow-up time for the high Ki-67 group was 64 months. Ki-67 might be a risk factor during long-term follow-up, while its effect is different in the early stages. Previous studies showed that high Ki-67 is related to more-invasive characteristics, such as perithyroidal soft-tissue invasion, vascular invasion, ETE, and distant metastases[35–37], meaning that early discomfort drives patients to seek and receive treatment promptly. In addition, since the Ki-67 range of DTC is generally < 10%[19, 38], higher Ki-67 in this range would probably not cause extreme outcomes.
Cancers with high proliferation tend to be more sensitive to radiotherapy. The possible reason for this is that radiotherapy preferentially eradicates rapidly dividing cells, while populations with slow proliferation show stronger resistance to radiotherapy[39]. Interestingly, an in vivo experiment indicated that low Ki-67 predicts high iodine affinity[10], but this trial was conducted at one single time point and in the absence of high TSH. Our current study found that patients with lower Ki-67 achieved worse prognoses. This might be the first discrepancy in TC to be found, it needs further verification and exploration in a larger cohort.
Most studies focused only on primary lesion’s Ki-67 expression, but once the tumor is removed, the antigen is biologically and functionally inconsistent with metastatic tumor cells and therefore no longer affect prognosis. Metastases grow in different immunological and trophic settings than the primary tumor, such as in lymph nodes, lungs, and bone, and these settings affect the morphological and institutional heterogeneity of the metastatic cancer cells. Nilsson et al. found that iodine avidity in LNMs was lower than in primary lesions because Tg gene mutations are more common in metastases[10]. Therefore, Ki-67 must be assessed in metastatic foci to determine how it affect clinical prognosis[40], which is a limitation of this study. Another limitation is that our samples lacked the analysis of PTC subtypes, WHO pointed out that PTC subtypes are also important factors affecting the prognosis, such as tall cell, columnar cell, and hobnail cell, etc., and it is not yet known whether they would have differences in Ki-67 values.