Introduction
Epithelial ovarian cancer (EOC) is the most lethal gynecologic malignancy and is closely associated with tumor recurrence and chemoresistance [
1]. Approximately 75% of patients are diagnosed at an advanced stage with widespread peritoneal lesions and frequent chemoresistance [
2]. Retrospective studies have demonstrated a strong association between the size of post-surgical residual tumor and prognosis [
3]. At present, the gold standard for management is primary debulking surgery followed by platinum-based adjuvant chemotherapy. However, optimal debulking can be achieved in only 30–60% of stage III/IV ovarian cancers [
4‐
7]. Neoadjuvant chemotherapy followed by interval debulking surgery has been proposed as an alternative to conventional PDS. These indications are for patients with a low probability of optimal cytoreduction or reversible contraindications [
8].
However, the role of NACT and PDS is still controversial. Several studies, including randomized controlled trials, have reported a significantly higher optimal debulking rate in patients who underwent NACT-IDS but a similar overall survival between NACT and PDS [
9‐
13]. Several retrospective studies regarding stage IV ovarian cancer patients have reported that the survival outcomes in patients who underwent PDS were lower compared to patients who underwent NACT [
14,
15]. In addition, several meta-analyses have suggested that NACT is associated with worst outcomes when compared to standard PDS followed by adjuvant chemotherapy [
16‐
19].
Additionally, emerging evidence from recent studies showed a higher incidence of platinum-resistance after NACT at first relapse compared to patients who received PDS [
20‐
22]. However, studies by Rauh-Hain [
23] and da Costa [
24] reported that the incidence of platinum-resistance after NACT at first relapse was similar, whereas the risk of platinum-resistance at second relapse was higher: Rauh-Hain’s (adjusted OR 4.06,
P = 0.001) and da Costa’s (adjusted HR 1.92,
P = 0.009).
This has led us to wonder if the controversial results were relevant to individual differences. This hypothesis may explain why a small subset of patients failed to respond well during NACT and could not achieve optimal debulking by IDS, and thus poor prognosis. Even for the majority of patients who attained optimal debulking after NACT, due to differences in chemosensitivity, patient outcomes varied. In addition, patients who demonstrated initial chemosensitivity with complete debulking would still relapse. Hence, we intend to retrospectively compare the chemoresistance and survival between NACT-IDS and PDS. And The clinical characteristics of patients who benefited from NACT were further evaluated.
Discussion
Optimal cytoreduction is a critical prognostic factor for prolonged survival, whether it is performed before or after chemotherapy. Recent studies have suggested that NACT have reduced morbidity after surgery and are more suitable for optimal debulking surgery [
9,
11‐
13,
27,
28]. NACT followed by IDS (NACT-IDS) is considered to be an alternative to conventional PDS for treating advanced ovarian cancer patients and expected to have better prognosis. With the increased use of NACT, debates have risen as to whether NACT-IDS offers any benefits.
In line with several studies, we observed a dramatic increase in optimal cytoreductive rates in patients treated with NACT-IDS compared to PDS. However no survival benefits were observed. The majority of studies have demonstrated similar progression free survival (PFS) and overall survival (OS) between patients treated with NACT and PDS [
9‐
13]. In our study, the patients’ baseline characteristics were well balanced. We found a higher rate of optimal cytoreduction for patients in the NACT group, but showed no improvement in PFS and OS compared to patients in the PDS group (median PFS: 19.9 months vs 25.4 months,
P = 0.371; median OS: 52.3 months vs 67.8 months,
P = 0.209). This observation is consistent with the results observed in the majority of previous studies.
An explanation for this is that chemoresistance in ovarian cancer may be induced by NACT, thus leading to a worse prognosis if recurrence occurs. Three retrospective trials demonstrated that patients receiving NACT-IDS had a higher risk of platinum-resistant recurrence compared to patients who underwent PDS [
20‐
22]. However, it still remains unclear whether NACT could induce platinum-resistance. Our study showed that patients in the NACT-IDS group had a higher incidence of platinum-resistance at first relapse compared to patients in the PDS group, and was confirmed using multivariate regression analysis (adjusted OR 2.837,
P = 0.002).
To explain this finding, one possible hypothesis is that the larger the volume of cancer present when chemotherapy is initiated, the higher the likelihood of development of mutations and chemoresistance [
29]. Large bulky tumors are often necrotic and hypoxic. Their poor blood supply does not often lend itself to maximal intravenous or intraperitoneal chemotherapy delivery. Comparatively, smaller tumors are well perfused and easier to target. At the time of initial treatment, both chemosensitive and chemoresistant cells are present in the patient. Primary surgery decreases the tumor burden for both types of cells and decreases the amount of cells that could spontaneously mutate to chemoresistant phenotypes, and thus to some extent, reduce chemoresistance. However, by initiating chemical debulking first using the NACT approach, tumor cells have more time to build resistance. Introducing IDS in the middle of the six to eight cycles of chemotherapy may have an effect in decreasing tumor burden. However, patients are more susceptible to develop mutations and acquire chemoresistance due to the exposure of a larger tumor burden to chemotherapy before IDS [
29].
In addition, our study demonstrated that residual disease was another risk factor that was associated with higher platinum-resistant recurrence risk (adjusted OR 2.575,
P = 0.013). This is consistent with Rauh-Hain [
23] and Lim [
27]’s study suggesting that NACT enriches for cancer stem cells that exists in residual disease, which eventually leads to chemoresistance. Although the diameter of metastatic tumors could be reduced using NACT, even to less than 1 cm, the residual tumor tissue may contain cancer stem cells responsible for chemoresistant recurrence [
20,
27,
30]. It has been reported that a small subset (5 to 12%) of patients show pathologic complete response (pCR) with no residual tumor cells after NACT and excellent prognosis [
31,
32]. Hence, the crucial goal for NACT is to achieve no residual tumor after IDS or even pCR. Research should focus on finding therapies that increase pCR in patients.
Additionally, our data demonstrated that PDS patients debulked to no RD had the longest PFS and OS. For PDS patients debulked to RD<1 cm or to RD ≥ 1 cm, their prognosis had no significant differences. Although patients who were debulked to no RD after NACT (NACT-R0) had a significantly lower PFS and OS compared to patients with no RD by PDS (PDS-R0), they were comparable to patients with RD < 1 cm at PDS (PDS-R1). In addition these patients showed a relatively better outcome compared to patients with RD ≥ 1 cm at PDS (PDS-R2). However for patients who were debulked to RD < 1 cm (NACT-R1) or ≥ 1 cm (NACT-R2) after NACT, their prognosis were no better than patients with RD ≥ 1 cm at PDS (PDS-R2) .
These results suggest that, R0 resection at PDS was associated with the best prognosis. Debulking to RD < 1 cm provides a smaller but still significant benefit for patients with PDS. For these patients, aggressive surgical procedures should be performed if the tumor could be safely resected to microscopic levels. For patients following NACT who ended up with RD after IDS, their prognosis was usually quite poor. They were not able to achieve optimal cytoreduction at PDS. If NACT could have some effect on minimizing tumors to a certain level, resectability should be considered. However, due to the large tumor burden, the majority of the patients developed chemoresistance and showed no improvement after NACT. If chemoresistance could be predicted in advanced, PDS may be a better choice. Effective alternatives such as second-line chemotherapy, immune-therapy, targeted therapy or relevant clinical trials should also be recommended for these patients.
Patients debulked to no RD after NACT failed to show any evident benefits on survival compared to patients with macroscopic RD at PDS. We further compared the survival curves for patients in the NACT group without RD using subgroups based on chemosensitivity. Our results suggested that, among patients with no RD after NACT, survival was significantly better if the patients were chemosensitive compared to patients with RD at PDS.
Hence it is critical to identify these patients who may benefit from NACT for the assessment of complete resectability and chemosensitivity. In terms of unresectability, our study showed that NACT may not achieve benefits. These included multisectional intestinal invasion or severe pelvic adhesion with diffuse dissemination of tumor nodules, or a contracted omentum with dense adhesion to surrounding organs. Several methods have been proposed in the literature for pre-surgical evaluation, including CA125 levels, radiologic methods, peritoneal cancer index (PCI), and laparoscopic scoring [
33‐
37]. The ASCO and SGO practice guidelines regarding the use of NACT do advocate laparoscopy as a tool to predict surgical resectability
. Relevant studies have suggested several laparoscopic and CT scoring systems that had certain predictive values for optimal cytoreduction. However limitations of over-or-underestimation were present in these methods [
33‐
37]. With the improvement of surgical techniques, several previous ‘unresectable’ tumors could now be removed meticulously. Amendments and validation for a revised scoring system is now needed. More reliable predictors for resectability are essential for classifying tumors preoperatively and should be the subject of further study. In the 2018 SGO annual meeting on women’s cancer, Dr. Beryl suggested a new prospective. If R0 was not attained, low volume disease confined to single anatomic locations (≤1 cm-SL) may be an alternative [
38]. The study showed that patients with RD ≤ 1 cm involving multiple anatomic locations (≤1 cm-ML) had similar outcomes to suboptimal debulked (RD>1 cm) patients [
38]. Moving beyond complete cytoreduction, low volume RD may be another option for consideration.
Apart from resectability, platinum resistance should be taken into consideration before the use of NACT for AOC patients. Recognition of chemoresistant properties based on initial tumor biology and molecular phenotype in advance is complicated. Studies have demonstrated that tumor burden, tumor biology and even adjuvant chemotherapy itself was associated with chemoresistance. Gene mutations, include TP53, BRCA1/2, BRAF, KRAS, β-catenin, PTEN et al. [
39,
40] have been considered to be associated with ovarian carcinogenesis and chemoresistance. In addition, the predictive value of cancer stem cell related markers, like CD44, ALDH1 have been demonstrated for chemoresistance [
30]. However, there is still no single gene marker that have showed specificity, sensitivity, accuracy and effectiveness for chemosensitive assessment. A gene panel consisting of multiple genes may have the potential for chemoresistant evaluation. Further studies concerning the possible mechanism for acquired platinum resistance are needed.
There were several limitations to our study. First, this was a retrospective study that inevitably had a selection bias, as a small number of patients were excluded due to incomplete clinical data or lost during follow-up. In addition, patient selection for NACT or PDS relied on gynecological estimation and imaging. Secondly, due to limited patient numbers and the relatively short period of follow-up, death or progression only happened in a few patients. A larger cohort prospective study is currently in progress to validate our findings.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.