PTC is a curable, treatable disease, but occasionally causes recurrences and cervical LNM. In an attempt to establish a reliable prognostic scoring system, most published studies have primarily considered clinical factors such as age and sex, tumor size, and extra thyroidal extensions. Younger patients less than 16 years old, patients older than 45 years, and patients with PTC subtypes such as tall-cell, columnar-cell, and diffuse sclerosing variant types are known to have higher risks for recurrences (Schlumberger, 1998; Beasley et al., 2002a) and additionally, bilateral LNM may be a risk factor for lung metastasis (Lee et al., 2011). The biological behaviors of well differentiated thyroid malignant nodules were not considered when selecting a treatment strategy. Furthermore, the way in which tumors spread via the lymphatic system is not fully understood.
Since PTC can metastasize via the lymphatic channel to a regional cervical lymph node (Shushanov et al., 2000; Yasuoka et al., 2005; Yu et al., 2005; Longatto-Filho et al., 2008;
Moriya et al., 2011), and increased LVD is assumed to be one of the most important parameters associated with the aggressive behavior of a malignant tumor (Longatto-Filho et al., 2008), we hypothesized that nodules with more lymphatic vessel systems would have an increased risk of LNM. We do not know whether the existing tumor induces lymphangiogenesis or invades peritumoral vessels so, in addition to D2-40, we also analyzed VEGFR-3. Vascular endothelial growth factor -C, known to be a key factor in lymphatic vessel sprout, is involved in the proliferation and migration of endothelial cells, increased
- 21 -
vascular permeability, and lymphangiogenesis(Byrne et al., 2005; Yamazaki and Morita, 2006) and its receptor, VEGFR-3, was shown to be essential to lymphatic vessel development in embryos and in the maintenance of lymphatic endothelial cells in adults (Shushanov et al., 2000). There are relatively few reports about LVD and the expression of Vascular endothelial growth factor -D in malignant thyroid nodules and the relationships of intratumoral LVD and peritumoral LVD with LNM have not been previously published, to the best of our knowledge. In this study, we quantitated both lymphatic markers intratumorally and peritumorally since published reports have shown contradicting results.
Some reports claim that there are relationships between increased intratumorallymphatics and an association with the spread of tumors to regional lymph nodes (Beasley et al., 2002b;
Hall et al., 2003; Giorgadze et al., 2005; Petersson et al., 2009), while, in contrast, there are also published data showing that there are no relationships between the intratumoral LVD related to LNM or distant metastasis in breast cancer (Britto et al., 2009). A study by Garcia et al. showed that peritumoral LVD was not correlated with LNM, along with other published reports showing that LVD is more strongly associated with a good prognosis (de la Torre et al., 2006; Garcia et al., 2010). This result showed contradicting results suggesting that higher LVD of PTC nodules shows significantly increased cervical LNM. In this study, intratumoral LVD as well as peritumoral LVD showed high correlations with cervical LNM making it an important factor indicating tumor aggressiveness. Few authors suggested that peritumoral LVD was related to higher capsular invasion (p=0.045), along with other studies, and that it is the peritumoral, not the intratumoral, LVD that is related with aggressive cancer behavior (Giorgadze et al., 2005; Gombos et al., 2005; Garcia et al., 2010). In an age-
- 22 -
and sex-adjusted univariate setting, both factors were highly associated with cervical LNM and when we analyzed these factors in a multivariate setting, both were independent factors among the tumor size, multiplicity, bilaterality, and extrathyroid extension. Thus, this data indicates that expression levels of VEGFR3 and D2-40 should be considered with respect to tumor aggressiveness and cervical LNM. Although lymphovascular invasion (LVI) seen with H-E staining did show statistically significant differences between the two groups, 52 cases (82.5%) in group 2 did not show LVI with H-E staining whereas a higher percentage of VEGFR-3 and more than 50% of intratumoral D2-40 stained vessels were visualized in group 2 nodules. We did not evaluate the LVI seen with these two markers and correlated with cervical LNM, but we quantitated the LVD with cervical LNM state because this study was initially designed to compare the LVD in the two groups with correlations to US features. Pathological features such as larger size, extra-thyroid extension, and positive immunostaining for D2-40 and VEGFR-3 to evaluate the LVD can be used as strong predictive prognostic markers and have the potential to play significant roles in pathological routine workup of thyroid cancer with cervical LNM.
Thyroid cancer with cervical LNM, showing strong VEGFR-3 staining and increased LVD with D2-40 staining did not show any correlations with assorted US features such as echogenicity, orientation, shape, margin, and calcifications. None of these factors showed any significant differences between the two study groups. Our hypothesis when designing this study was that since water or vessels can show low echogenicity compared to solid mass comprised of compact cells, a tumor with higher LVD would be correlated with low or marked echogenicity, but the observed effect was not statistically significant. We also
- 23 -
hypothesized that spiculation or ill defined margins would be seen more often with an aggressive tumor with increased peritumoral LVD invasion, but the results for this factor were also not significant. However, Ito et al. reported that ill defined edges were positively associated with a more aggressive thyroid cancer frequently showing lateral LNM (Ito et al., 2005) and that the feature may be used as a marker of a more aggressive cancer. In contrast, in this study we observed that nodules with spiculation or irregular shapes occurred more frequently in group 1 with no statistical significance. This findings show that US features vary according to the size of the malignant nodule, rather than the aggressiveness of the tumor, showing more frequent “taller than wide” appearance in microcarcinoma (p=0.010) and more associated calcifications in PTCs larger than 1 cm. We attempted to identify US features associated with aggressive malignant features but none of the features examined were associated with cervical LNM and were, instead, more strongly associated with features of microcarcinoma. Calcifications were frequently involved in aggressive thyroid cancers with moderate staining of the lymphatic vessels.
One limitation of this study is that we analyzed clinical and pathological data retrospectively. We used age- and sex-matched pair data in order to concentrate on pathological factors and it was not possible to perform staining procedures for all the surgical specimens. Furthermore, it is possible that a considerable amount of selection bias may have affected this study by comparing patients with no cervical LNM in the final pathology group with a cervical LNM-positive group that included both central and lateral LNM-positive patients. We also excluded patients with central LNM because a considerable number of cases showed micrometastasis to the central lymph nodes and such cases have
- 24 -
little clinical impact due to the routine surgical removal of the ipsilateral central lymph nodes, but it may be considered as another form of selection bias. We did not follow up the patients with high risk factors to track recurrences since this study was not designed to use a prospective cohort and it was not focused on how LVD relates to survival, but specifically how it is related to cervical LNM state. In the future, it would be worthwhile to investigate the relationship between LVD and patient survival. Quantitation of intratumoral LVD and peritumoral LVD by a counting method revealed an association (P= 0.027) between high intratumoral LVD (highest tertile) and LNM in PTC, but in this experimental setting, we were unable to conclude if this was newly developed lymphangiogenesis or if existing lymphatic vessel density caused this result. To overcome this problem VEGFR-3 was also stained to confirm the presence of existing lymphatic channels. Quantifications were obtained but due to the generally weak, moderate, or strong nature of the staining for VEGFR-3, it was difficult to compare the association between the existing vasculature and newly developed lymphangiogenesis.
- 25 -