1. Introduction
Surgical treatment for cervical cancer (CC) has been discussed intensively owing to recent advances in minimally invasive surgery (MIS). The technical feasibility of minimally invasive radical hysterectomy (MIS-RH) has been previously described in numerous reports [
1,
2,
3,
4,
5], but there are increasing concerns regarding the oncological outcomes, especially on the basis of the results of the recent prospective randomized trial, i.e., the international laparoscopic approach to cervical cancer (LACC) trial [
6,
7,
8,
9,
10]. Conventional abdominal RH (A-RH) showed better disease-free survival (DFS) and overall survival (OS) than MIS-RH [
8,
9], but the current evidence is still debatable [
11,
12,
13,
14,
15]. The reason is probably because MIS-RH includes both conventionally laparoscopic RH (L-RH) and robotic RH (R-RH), and the techniques needed for conventional laparoscopic surgery are much more advanced and complicated than those for robotic surgery. L-RH requires more time for surgeons to develop the essential level of skill to perform the procedure effectively and has a longer learning curve. Many studies have compared the feasibility of these two different MIS techniques for treating different types of gynecological diseases, regardless of the malignancy status. Unsurprisingly, the results favor the robotic approach [
16,
17,
18,
19,
20,
21,
22]. Therefore, when most MIS-RH procedures in the LACC trial were performed via conventional L-RH (84%) and contributed to poorer survival outcomes [
9], one could argue that L-RH is not regularly performed in most cancer centers and the current practice and trend favor R-RH in many countries [
12], which is also apparent in Taiwan [
23,
24].
It is also debatable whether RH can be performed instead of concurrent chemoradiotherapy (CCRT) for the management of locally advanced CC. In addition to the well-known criteria such as FIGO (International Federation of Gynecology and Obstetrics) staging, tumor size is among the most concerning factors with an increased possibility of therapeutic failure, and subsequently decreased DFS and OS, regardless of whether surgery was performed [
25,
26,
27,
28,
29]. Many strategies have been suggested to overcome this limitation, and neoadjuvant chemotherapy (NACT) has been introduced for treating various types of malignancies [
29,
30,
31,
32,
33,
34] and locally advanced CC [
28,
29,
30]. The main purpose of NACT is to reduce tumor size, thereby facilitating subsequent local treatment. Among patients with CC, those who received NACT for FIGO stage IB2/IIA2 CC were highly likely to undergo successful laparoscopic surgery [
30]. Other advantages of NACT include the potential of controlling micro-metastatic disease and reducing the need for postoperative adjuvant radiotherapy with/without chemotherapy (CT) [
35,
36]. To date, only a few studies have evaluated the feasibility and safety of MIS-RH for patients with CC after NACT [
37,
38,
39,
40,
41,
42,
43]. In our previous study, we evaluated the therapeutic effect of RH after NACT in patients with locally bulky-size cervical cancer (LBS-CC ≥ 6 cm). Weekly NACT with only cisplatin significantly decreased tumor size (from 6.4 ± 0.5 cm to 4.5 ± 1.4 cm), reduced blood loss (from 930 ± 356 mL to 558 ± 1328 mL,
p < 0.001), and lowered the immediate complication rate (from 32% to 5.7%) during subsequent A-RH [
28]. A meta-analysis of 18 studies with 1785 patients with CC (mainly FIGO stage IB2 to IIB) who were treated with cisplatin-based NACT before RH confirmed that the response to NACT was an indicator of progression-free survival (PFS) and OS [
35], suggesting that these patients would have better oncological outcomes if the tumors showed a good response to NACT treatment. Therefore, in this retrospective study, we evaluated and compared the surgical and oncological outcomes of conventional A-RH and R-RH in patients with LBS-CC after NACT.
4. Discussion
The results of the current study demonstrated that NACT-R-RH resulted in significantly better peri-operative outcomes than NACT-A-RH did, with a shorter operative time, lower estimated blood loss, fewer incidences of intraoperative blood transfusion, lower immediate postoperative and 24-h VAS scores, and shorter LOS, without increasing intraoperative complications; these results strongly indicate the feasibility of R-RH in patients with CC. This observation is not new, as a recent meta-analysis suggested that R-RH had surgical benefits for patients with CC than A-RH did [
22]. Although the perioperative benefits of R-RH have been well documented, most studies were performed for early stage CC, regardless of whether patients received NACT treatment [
12,
16,
17,
18,
19,
20,
21,
22,
37,
38,
39,
40,
41,
42,
43]. Some studies did not support the benefits of NACT in patients who were scheduled to undergo surgery, especially patients needing MIS [
48,
49,
50]. One of the frequent observations was the desmoplastic effect of the tissue after NACT, making tissue dissection much more difficult during surgery [
48,
49,
50]. In the current study, although 2 patients in the NACT-R-RH group had bladder injury, compared to none in the NACT-A-RH group, we do not believe that this complication is due to the desmoplastic effect of NACT, because all patients in the NACT-R-RH group underwent complete R-RH and no conversion to laparotomy was necessary. Furthermore, intra- and postoperative parameters, such as operative time, estimated blood loss, the need for blood transfusion, and LOS, were significantly better in the NACT-R-RH group than in the NACT-A-RH group; this finding was similar to the findings of previous studies that showed the benefits of NACT, without increasing the difficulty in performing MIS techniques [
40,
42,
51]. In another study, the median estimated blood loss, operative time, and LOS were significantly better after NACT-R-RH than after NACT-A-RH [
51]. In a previous study, although the mean operative time was longer in the NACT-R-RH group than in the NACT-A-RH group, other intra-operative parameters, such as estimated blood loss and LOS, were still significantly better in the NACT-R-RH group [
50]. In both studies, no significant difference was observed in the recurrence pattern, DFS, and OS during the 3-year follow-up between NACT-R-RH and NACT-A-RH [
50,
51]. A recent review of many previous case-control and retrospective case series also demonstrated the safety of NACT-R-RH considering the oncological outcomes [
52].
The goal of NACT includes downstaging the tumor to improve the radical curability and safety of surgery and inhibit micro-metastasis and distant metastasis [
29,
30,
31,
32,
33,
34,
35,
36], as well as allow surgeons to perform less radical surgery without compromising oncological safety [
52,
53]. An early Cochrane review showed a significant decrease in adverse pathological findings after NACT (odds ratio [OR] 0.54, 95% CI 0.40–0.73,
p < 0.0001 for LN status; OR 0.58, 95% CI 0.41–0.82,
p = 0.002 for parametrial infiltration), contributing to an improvement in both OS and PFS. These results appear to indicate that NACT may offer benefits over surgery alone for women with LBS-CC, although the effects were less clear for all other pre-specified outcomes [
53]. A recent meta-analysis also supported the benefits of decreasing the severity of disease status in patients with LBS-CC treated with NACT, based on the significantly decreased rate of LN metastasis, parametrial infiltration, and recurrence [
54]. Thus, NACT is an acceptable and effective procedure for selected patients with LBS-CC, although the analysis did not support advantages in survival (DFS, PFS, and OS) of patients with LBS-CC treated with NACT [
54]. In fact, the impact on survival after NACT is still highly debatable, with no consensus about whether it can result in better outcomes [
28,
29,
30,
35,
36,
54,
55,
56,
57,
58,
59]. Accordingly, the term “acquired treatment response” seems important. A meta-analysis showed varying response rates after NACT for patients with LBS-CC, ranging from 58.5% to 86.5% for the clinical response rate and from 7.5% to 78.8% for the pathological response rate [
57]. The clinical response rate predicted favorable DFS (HR 2.36, 95% CI 1.82–3.06) and OS (HR 3.36, 95% CI 2.41–4.69) [
57]. Moreover, the pathological response also predicted favorable outcomes such as DFS (HR 3.61, 95% CI 2.0–6.52) and OS (HR 5.45, 95% CI 3.42–8.7) [
57]. Another meta-analysis further suggested the importance of an early response, because an early response to NACT is associated with favorable outcomes [
58]. However, it is difficult to define “early response”. In the current study, some adverse surgical and pathological parameters seemed to be improved after NACT, although the clinical relevance is unclear. We did not use the same evaluation tools to compare the changes before and after treatment. For example, in the current study, the tumor size decreased by approximately 25% (from 4.6 cm to 3.4 cm in the NACT-R-RH group and from 5.1 cm to 3.6 cm in the NACT-A-RH group, respectively). The size before treatment was calculated using preoperative imaging but the size after treatment was based on postoperative pathological measurement. It is well known that formalin-fixed tissue is significantly smaller than tissue before the process of formalin fixation, with the percentage of shrinkage being approximately 25–30% [
28]; therefore, in the current study, tumor size showed almost no response, which might partly explain the worse outcomes.
Another issue in the current study is the use of different NACT regimens for the patients with LBS-CC. In the NACT-R-RH group, more patients received NACT with taxane/platinum-based regimens. It is unclear whether the different NACT regimens contributed to varied outcome in our study. In another study, among women with LBS-CC treated with NACT-RH, taxane/platinum-based regimens had a similar effect on DFS and cause-specific survival compared to non-taxane/platinum-based regimens, irrespectively of the tumor type (squamous cell carcinoma and non-squamous cell carcinoma) [
55]. In the current study, the NACT regimen administered was significantly different between the NACT-R-RH and NACT-A-RH groups, and the taxane/platinum-based regimens were administered much more frequently in the NACT-R-RH group. The results from univariate analysis showed that the taxane/platinum-based regimen was associated with worse prognosis, but it may have been influenced by other confounding factors. Further multivariate analysis was performed, and the results did not establish any relationship between different NACT regimens and prognosis. Hence, as the evidence about the beneficial roles of NACT is still not conclusive, further study is warranted.
The worse prognosis of patients with LBS-CC treated with NACT-R-RH demonstrated in the current study led us to re-visit recent publications in the New England Journal of Medicine [
8,
9]. The LACC trial showed a decreased OS and a 3-fold increase in recurrence in the MIS-RH group than in the A-RH group, both of which were unexpected results [
9], suggesting that MIS-RH should be used cautiously. In the current study, FIGO stage IIB was the most important indicator for worse outcomes of patients with LBS-CC treated with NACT-RH. Per the new FIGO staging of cancer of the cervix uteri (2018), invasive carcinoma limited to the cervix with the greatest dimension of ≥2 cm and <4 cm is classified as stage IB2, carcinoma >4 cm as stage IB3, and carcinoma that is limited to the upper two-thirds of the vagina without parametrial invasion and >4 cm as stage IIA2 [
25]. This new FIGO staging system of CC emphasizes the importance of tumor size. In a previous study, Wright et al. used the National Cancer Database (
n = 62,212, data used were of patients treated between 2004 and 2015) to examine the prognostic performance of the 2018 CC staging system; the 5-year survival rate was 91.6% (95% CI 90.4–92.6%) for stage IB1 CC, 83.3% (95% CI 81.8–84.8%) for stage IB2 CC, and 76.1% (95% CI 74.3–77.8%) for IB3 CC [
60]. Another group used the data of The Surveillance, Epidemiology, and End Results Program collected between 1988 and 2014 (
n = 8909) to validate this new FIGO CC staging system and also confirmed that this system was valuable for distinguishing survival outcomes of patients with FIGO stage IB CC [
61]. In that study, the HR of cause-specific survival of 2018 FIGO stage IB3 CC and 2018 FIGO stage IB2 CC was 4.07 (95% CI 3.33–4.97) and 1.98 (95% CI 1.62–2.41), respectively, compared to that of 2018 FIGO stage IB1 CC [
61]. On multivariable analysis for cause-specific survival in the FIGO stage IB cohort based on 2018 FIGO stage IB2, the HR for 2018 FIGO stage IB3 was 2.06 (95% CI 1.76–2.41) and HR for 2018 FIGO stage IB1 was 0.51 (95% CI 0.42–0.95) [
61]. Thus, the 2018 FIGO CC staging system reflected the effect of screening and prevention programs for CC [
62,
63,
64,
65,
66,
67], as the proportion of 2018 FIGO stage IB1 (
n = 3604, 40.5%) and stage IB2 disease (
n = 3620, 40.6%) was similar while the percentage of 2018 FIGO stage IB3 disease was only 18.9% (
n = 1685) [
61]. It was anticipated that researchers would find trends in the significantly increasing proportion of 2018 FIGO stage IB1 CC and in the decrease in the 2018 FIGO stage IB3 disease during the study period [
61]. The data presented in the current study were consistent with the new 2018 FIGO staging system for CC, and we re-considered the treatment strategy for patients with LBS-CC. Although the cut-off value of 2 cm (2018 FIGO stage IB1 CC) might be applicable and RH could be performed as the MIS procedure [
68,
69], this cut-off is still debatable [
8,
9]. Therefore, we strongly agree with the notice of the Society of Gynecologic Oncology (SGO) announced in November 2018, that gynecological oncologists should be aware of the emerging data on MIS for CC so that “thorough discussions can be undertaken with patients and shared decision-making can be used when choosing the surgical approach for RH”. Moreover, the results of the LACC trial, together with institutional data, should be discussed with patients before choosing MIS-RH.
To address the potential risks of MIS for gynecological cancer, many researchers have suggested and hypothesized possible mechanisms that contribute to worse prognosis, thereby assisting in minimizing tumor dissemination during surgery [
70,
71,
72,
73,
74,
75,
76,
77,
78]. Most hypotheses are mostly inconclusive and require larger randomized controlled trials. In addition, MIS-RH is indeed a challenging surgical procedure compared with conventional total hysterectomy [
78]. As the surgical technique for MIS-RH varies greatly and is extremely difficult to control for with randomization or multivariate adjustment [
69], good training to perform the technique delicately is of paramount importance [
79,
80,
81,
82,
83].
Several limitations of the current study are apparent due to the unfair comparison between two groups. The first limitation is the small sample size of both groups. The wide use of effective universal screening for cervical cancer (pap smear) in developed countries [
62,
84,
85,
86,
87] has led to a dramatic decrease in LBS-CC prevalence rate. Moreover, the therapeutic choice for patients with LBS-CC is debated [
88]; therefore the patients enrolled in the current study were mainly based on shared decision making, which directly influenced the sample size, resulting in an insufficient power to properly compare PFS and OS between the two types of surgery after NACT. Second, based on the same reason shown above, an additional cost is required for receiving robotic surgery, which is not covered by the National Insurance Health Care System in Taiwan [
89,
90,
91,
92,
93,
94], and the risk of selection bias could not be totally avoided, contributing to a significantly greater number of patients with FIGO stage IIB in the robotic group. This is a well-known limitation of the retrospective design. The effects of both limitations were combined and augmented by the confounding effects. Finally, imaging evaluation was not used at the end of NACT treatment. Therefore, the response of the patients to NACT was uncertain.
Despite these limitations, the current study provided valuable data showing that women with LBS-CC treated with NACT followed by surgery required further evaluation, especially for those patients with FIGO stage IIB. Our study confirmed that the seemingly promising perioperative outcomes in the NACT-R-RH group did not translate into better oncological outcomes. In contrast, they were associated with higher recurrence and mortality rates. Therefore, it is clear that any advanced technology (robotic surgery) in cancer treatment should be testified by survival outcomes, regardless of perioperative outcomes. In addition, the current study provided additional evidence that FIGO stage IIB contributed greatly to a poorer prognosis if NACT-R-RH was performed. Therefore, it is reasonable to suppose that NACT-A-RH might be a more appropriate approach in patients with FIGO stage IB3-IIA2 LBS-CC if the patients plan to undergo NACT along with a subsequent surgical intervention.