To our knowledge, optimal tumor debulking and platinum response have been proven to be the most powerful prognostic factor for both overall survival and progression-free survival of ovarian cancer patients. To acquire optimal tumor debulking, multivisceral resection including en bloc resections with bowel resection, upper abdominal procedures, and extensive peritonectomy are required when necessary. Despite improvement in surgical techniques, there are still patients who are not obtain radical resection will develop platinum-resistance with poor prognosis. At present, CA125 are conventionally used in monitoring responds to surgery and chemotherapy for ovarian cancer patients. Data from GOG-182 showed that the median OS for patients whose CA125 value declined to normal level after 2nd cycle chemotherapy was 77.7 months, compared with 23.0 months for those who did not normalized, and improved PFS was observed in patients with CA125 value declined to normal level after 1st, 2nd or 3rd cycle chemotherapy compared to those who never normalized before 4th cycle chemotherapy [14].
HE4 is a novel tumor biomarker in EOC patients. The assessment of the prognostic significance of pretreatment HE4 has been described in many papers. However, few studies focused on the early clearance of HE4 after treatment in predicting prognosis of epithelial ovarian cancer. There was some result suggested that HE4 was mainly secreted by malignant ovarian cancer cell and tumor micro-environment [15, 16]. Therefore, the removal of the tumor and the response to treatment should correlate with the clearance of HE4.
A prospective study of Roberto Angioli suggested that HE4 reduction with above 47% at the third cycle of chemotherapy were more likely to be platinum sensitive. On the contrary, CA125 value did not correlate with platinum response [17]. In our recent study, we have demonstrated that single HE4 superior to CA125 in predicting platinum sensitivity. Our data showed that the maximum AUC value of HE4 alone was 0.779 (p = 0.000) in predicting platinum sensitivity after the third cycle chemotherapy compared to the maximum AUC = 0.731 of CA125 (p = 0.004) after 1st cycle chemotherapy respectively. When the two biomarkers were combined, the result showed that when HE4 clearance after 3rd cycle chemotherapy or CA125 clearance after 1st cycle chemotherapy, the AUC, sensitivity and specificity were 0.788, 100% and 57.5% respectively. It means that 100% patients with platinum resistant could be identified through the both non-clearance of HE4 after 3rd cycle chemotherapy and CA125 after 1st cycle chemotherapy. When HE4 after 3rd cycle chemotherapy and CA125 after 1st cycle chemotherapy were both clearance, 94.5% individual with platinum sensitive could be recognized. The similar finding is in line with earlier studies [17–20]. Anita et al. have reported that the predictive abilities with regard to platinum sensitivity for HE4 was excellent as the AUC = 0.846, closely with our result [19]. Besides, Chen et al. concluded that the change of HE4 was more closely related to the chemotherapy response compared to the change of CA125 [21]. These results suggest that monitoring HE4 and CA125 during chemotherapy should be recommended. It may help early identifying high-risk patients with platinum-resistant in ovarian cancer. In other words, for this part of high-risk patients with a slow decline in HE4, the corresponding imaging evaluation should be developed. At present, there are no clinical trials showing that modifying the treatment based on the unsatisfactory decline of serological tumor biomarkers after treatment can improve the outcome of patients. HE4 and CA125 appear to able to select high-risk patients who demand further treatment base on their adverse features, but still need to be confirmed in larger studies.
Olivier Colomban et al. used a Kelim model [22], which characterize the CA125 elimination rate during the first 100 days of neoadjuvant chemotherapy (NACT) and adjuvant chemotherapy, to assess the benefit in survival with bevacizumab addition for high-risk ovarian cancer patients in ICON-7 [23]. The result showed that only those high-risk patients with an unfavorable KELIM parameter less than 1.0 might have derived a benefit from bevacizumab when considering non-censored median survivals. With respect to HE4, previous study showed that patients with HE4 change of > 80% during NACT in advance high-grade serous ovarian cancer correlated with prolonged OS compared to change < 80% [24]. However, serum CA125 decline of > 80% and < 80% during NACT had no statistical significance in OS. Patients with CA125 logarithmic decrease or normalization within one month post-operative were correlated with better PFS and OS. Our data also demonstrated that serum HE4 is a more promising biomarker in prognosis of EOC compared to CA125. In the ROC curve analysis, patients with HE4 level normalization or reduction above 90% after 3rd and 6th cycle chemotherapy significantly correlated with two-year PFS (AUC = 0.707, p = 0.001, and AUC = 0.666, p = 0.011). CA125 level normalization or reduction above 90% after 1st cycle chemotherapy was correlated with two-year PFS, with a AUC of 0.648, p = 0.023. When combined with HE4 after 3rd cycle chemotherapy and CA125 after 1st cycle chemotherapy, it was shown that the AUC reached to 0.730 when one of them declined above 90% or normalized (p = 0.000), reporting an 83.3% of sensitivity and a 62.7% of specificity. Approximately 83.3% of patient who relapsed or progressed in two years could be identified during first-line chemotherapy with a 62.7% of specificity based on the combination of HE4 and CA125 early clearance. Using biomarkers to identify high-risk patients may complement the current definition of high-risk patients.
The result of Kaplan-Meier survival curve and long rank test also demonstrated that both the clearance of HE4 after 3rd cycle chemotherapy and CA125 after 1st cycle chemotherapy were significantly correlation with the PFS and OS. Prolonged PFS was significantly impacted by the pretreatment HE4 value but not pretreatment CA125. As respected to OS, there were no significant correlation between pretreatment value of HE4 and CA125 in our results. Similar results were obtained in other research which analyzed the influence of the pretreatment HE4 on PFS [8, 14, 25, 26]. Amanda et al. analyzed a group of 3686 patients with ovarian cancer and demonstrated that there was no difference in pretreatment CA125 with outcome [14]. However, patients with CA125 that normalized after 1st, 2nd, 3rd cycle of chemotherapy treatment were less likely to experience disease progression as compared to those of not normalization ones. Anita et al. analyzed a group of 48 EOC patients treated with primary debulking surgery and demonstrated that the prolongation of PFS and OS was significantly correlated with the pre-operative HE4 value [19]. Their result also concluded that the duration of OS was significantly influenced by the HE4 value after the third course of chemotherapy but not CA125, and the prolonged PFS was influenced by the CA125 value after the third course of chemotherapy treatment bot not the HE4. This result was different from ours. The reason for this problem may be due to its small sample size and selection bias.
This study was a retrospective character, single-center setting, relatively small number of patients, and lack of external validation, which may bias the results. Including more clinical covariate such as ECOG score, oncogene expression might be helpful to provide more information. Therefore, further studies are needed to confirm these results.
In conclusions, HE4 is a more promising biomarker in epithelial ovarian cancer for predicting prognosis and judging platinum sensitive compared to CA125. The early clearance of HE4 and CA125 were related to the PFS and OS. Monitoring the dynamic value of both HE4 and CA125 during treatment might be helpful in predicting platinum sensitive and risk to progress and relapse.