Significance of serum and pathological biomarkers in fertility-sparing treatment for endometrial cancer or atypical hyperplasia: a retrospective cohort study

The incidence of Endometrial cancer (EC) is rising in China in recent years, with the increasing prevalence of metabolic syndrome. As young patients tend to postpone delivery, more patients will face the contradictory situation of cancer treatment and uterus preservation. Many studies have proven the efficacy and safety of progestin-based fertility-sparing therapy for patients with EC or endometrial atypical hyperplasia (EAH), with a relatively high neoplasm remission rate of 75.3–88.7% [1].

Excessive unopposed exposure of endometrium to estrogen has long been regarded as the most important risk factors for endometriod cancer [2]. Additional risk factors include abnormal glycometabolism, insulin resistance (IR) and thyroid disease [3‐5]. But the influence of such serum biomarkers on treatment efficacy hasn't been fully discussed under the fertility-sparing situation. Well-differentiated EC also pathologically presents with positive expression of estrogen receptor (ER) and progesterone receptor (PR) associated with good prognostic [6]. But it's unclear how such pathological biomarkers change during the conservative treatment. We neither know whether their expression or changed status relates with therapeutic response.

This study analyzed the dynamic change of serum and pathological biomarkers and explored their relationship with treatment outcomes in order to find efficacy predictive markers for fertility-sparing therapy of EC or EAH.

The objective of this study was to determine the changes in expression of serum and pathological biomarkers during progestin therapy for EAH and EC patients and what implications in prediction of treatment efficacy would be expected.

Our study found, no significant change in HbA1c or HOMA-IR during fertility-sparing treatment, which implies such a dose of progestin does not aggravate or alleviate the status of insulin resistant or glycometabolism for patients. Our results also indicated that a higher baseline level of HOMA-IR related to a longer remission time. However, for patients with longer remission time their HOMA-IR level tended to decrease following treatment than those with faster remission (− 44% vs. 25%, p = 0.020). This might be due to the combined use of metformin and weight loss, as often suggested by doctors for patients who had initial insulin resistance. The result also proved that more patients in this group used metformin (41.9% vs. 26.3%) and their status of insulin resistance has an expectant improvement.

Further multivariate Cox analysis confirmed that patients with HOMA-IR ≥ 2.5 correlated with longer remission time. Insulin resistance has long been regarded as an important risk factor for the development of EC [9, 10]. However, few reports have analyzed the role of IR in fertility-sparing treatment. A review showed diabetes mellitus (DM) seems not affect the outcome of conservative treatment in EAH and EC, though the number of included patients diagnosed with DM is small [11]. Yang [12] analyzed EAH patients and found that, IR significantly affected the treatment time to remission (8.1 months with IR vs. 6.1 months without IR, p = 0.004). Our result also proved that for EC and EAH patients, IR status prolonged the treatment time. We propose IR might negatively affect the treatment response to progestin through several mechanisms. Firstly, insulin could act directly on EC cells through the PI3K/AKT/mTOR signaling pathway that promotes cell survival and proliferation [13‐15]. Secondly, insulin resistance is described to suppress hepatic sex hormone-binding globulin production and induce ovarian steroid hormone production which result in increased bioavailable estrogen and alleviates the antiestrogen effect [16]. Additionally, insulin could induce cholesterol synthetase DHCR24 expression which inversely correlates with PR and thus aggravating progestin resistance in EC cells [17]. Therefore, we believe insulin resistance would compromise the therapeutic effect of progestin and thus prolong the treatment time. For such patients with IR, it's reasonable to treat them more than 6 months or even longer. On the other hand, we should consider the benefit of the combined use of metformin to improve their IR status which might potentially improve the therapeutic efficacy.

Regarding the pathological markers, we found that the levels of PR and its subtype PRB diminished post-treatment, and patients with marked downregulated levels of PRB at the 3–6 months of treatment predicted faster remission. This might be because a downregulation of PRB in good responders could be indicative of active receptors which is usually ubiquitinated and turned over rapidly. For now, reports regarding the utility of hormone receptor expression in predicting the response to progestin-based therapy have been controversial [18]. In consistent with our finding, Vereide’s research in endometrial hyperplasia (EH) patients showed that PR, ER or their isoforms were downregulated in responders and stably expressed in nonresponders [19]. Raffone's retrospective study showed a weak stromal PRB expression is a highly sensitive predictive marker of no response of AEH and EEC conservatively treated [20]. Another small study of a 10-patient cohort found that ER and PRB levels after treatment with a progestin-containing IUD were significantly higher in the “progression” group than in the “no progression” group [21]. Also, a systematic review showed PR is significantly predictive of response in EH and EC treated with a levonorgestrel-intrauterine device, but not with oral progestins [22]. However, other studies did not report any significant associations [23, 24]. The GOG211 trial found that only low pre-treatment levels of the Ki-67 level were a predictor of the histologic response, rather than ER or PR [25]. Given the high remission rate to progestin therapy, the treatment time rather than the remission rate might contribute to above inconsistent results. Additionally, the small sample size of the above studies might have prevented a reliable conclusion to be drawn. Our results imply that greater change in PRB expression after initial treatment may serve as promising markers for early predicting response to conservative therapy.

There are several limitations to our study. First, retrospective data collection might produce bias with information ascertainment. Second, our relatively small sample size reduces the ability to draw more convincing conclusions. A large-sample prospective study is needed to verify our preliminary findings. Third, due to the limited number of patients, we could not investigate the effect of metformin on changes of serum or pathological biomarkers. This implication remains to be further demonstrated. However, this study described a comprehensive picture for the first time of changing trend in both endocrine hormone levels and pathological markers during fertility-sparing treatment. This study also represents one of the largest series in the literature which proved the significance of baseline insulin resistance and the marked change in PRB proteins in predicting treatment efficacy.

In conclusion, the current study shows that during fertility-sparing treatment for EC and EAH patients, insulin resistance status or thyroid function is not affected, while progesterone receptor level diminishes significantly. As for treatment efficacy, insulin resistant status of patients predicts longer treatment time to remission, while marked reduction of PRB after 3–6 months' treatment has good implications for faster remission.