3.2. Patient Characteristics and Treatment Modalities
The total study population of this review consists of 7544 women with cervical cancer, 3944 with epithelial ovarian cancer, and 1229 with endometrial cancer. The median ages are 31.6, 29.0, and 32.6 years, respectively. The histology, stage, and treatment modalities per tumor type are shown in Table 1
, Table 2
and Table 3
In patients with cervical cancer, squamous cell carcinoma (SCC) was the most common subtype. The majority of patients were diagnosed with stage IB1 (FIGO 2009) disease. FSS consisted of simple or radical trachelectomy, conization or large loop excision of the transformation zone (LLETZ), with or without pelvic lymphadenectomy. In 12 studies, patients received NACT prior to FSS. The majority of patients underwent an LLETZ, conization, or simple trachelectomy (ST). In total, 3.9% of patients (307/7851) initially planned for FSS had positive lymph nodes, positive resection margins, or other reasons contraindicating FSS. These patients underwent non-FSS treatment. In 2018, the FIGO staging for cervical cancer was revised [169
]. All but one study ([26
= 32) in this review used the former FIGO 2009 classification.
The majority of patients with epithelial ovarian cancer had a low-grade tumor (76.7%), mucinous histology (45%), and was diagnosed with stage IA or IB (52.6%). FSS consisted of unilateral salpingo-oophorectomy and preservation of the uterus and contralateral ovary, with or without additional staging procedures. In 18 studies, the completeness of surgical staging was specified. Complete peritoneal staging including peritoneal washings and/or biopsies and omentectomy ranged from 41 to 100% with an average of 70% (776/1110 patients). Pelvic and/or para-aortic lymph node sampling was performed in 56% (617/1110 patients). Of the patients with known details about the surgical technique, 275 of 426 (64.6%) underwent a laparotomy and 151 of 426 (35.4%) a laparoscopy. In total, 1815 of 3586 patients (50.6%) received adjuvant chemotherapy.
Almost all patients with endometrial cancer were diagnosed with grade 1 (92.4%), endometrioid adenocarcinoma (90.3%), and stage IA (83%). In 14 studies, the type of FSS prior to the start of hormonal therapy was reported and consisted of hysteroscopic resection (HR) and/or dilatation and curettage (D&C). In all 23 included studies, patients received hormonal therapy with progestin, although type, dosage, and treatment scheme differed considerably between studies. The majority of patients were treated with medroxyprogesterone acetate (MPA) or megestrol acetate (MA). MPA was used in 14 studies, all with different dosages, ranging from 20 to 1500 mg/day. In 17 studies, MA was used, in most patients with a dosage of 160 mg/day (range 40–800 mg/day). Patients received a levonorgestrel intrauterine device (LNG-IUD, 20 or 52 mg) as a single treatment in 5 studies (n = 91) and it was used in combination with oral progestin in 6 studies (n = 45). In 2 studies, metformin was given in addition to oral progestin and in another 2 studies, progestin was combined with GnRH analogues. The median duration of hormonal therapy with oral progestin ranged from 4.2 to 10.9 months and with LNG-IUD from 9.1 to 25.8 months.
3.3. Oncological Outcome
In Table 4
, the oncological outcome after FSS in each of the gynecological malignancies is presented. Not all studies reported both recurrence and death from disease.
In 29 studies on cervical cancer, involving 2432 patients, LLETZ, conization, or ST with or without pelvic lymphadenectomy was performed. In total, 28 (3.6%) patients had recurrent disease and 6 (0.8%) died of disease during a median follow-up of 53 months (range 9–131 months). More detailed information was available of 23 recurrences. Significantly more recurrences occurred in patients with FIGO stage IB1 (3.1%) and IB2 (5.6%) disease than in patients with FIGO stage IA1 (0.2%) and IA2 (0.7%) disease. SCC recurred in 0.7% of patients, adenocarcinomas (AC) in 1.1%, and adenosquamous cell carcinomas (ASCC) in 4%. The differences between these histological subtypes are statistically not significant (p = 0.159).
In 31 studies, including 2401 patients, a vaginal radical trachelectomy (VRT) was performed. Recurrent disease occurred in 82 (4.2%) patients and 34 (1.7%) patients died, after a median follow-up of 52 months (range 9–131 months). Unfortunately, several studies did not report on tumor size and LVSI, or combined stage IA and IB1 disease, which are factors associated with recurrent disease. In 7 studies (n = 468 patients) reporting on recurrence rates, details on initial tumor size were described. The overall recurrence rate was 6.2%, divided into 3.5% in patients with tumors <2 cm and 20.5% in ≥2 cm (p ≤ 0.001). None of the patients with tumors <2 cm received NACT, although 5 of 83 patients with a tumor size ≥2 cm received NACT. In 12 studies (n = 1134), detailed information on recurrence according to histological subtype was described. The overall recurrence rate was 4.1%. In patients with SCC or AC, it was 3.4% and 6.1%, respectively (p = 0.03). In only 5 studies (n = 656), detailed information on initial LVSI in patients who had recurrent disease was reported. LVSI was found in 31.3% of patients. The overall recurrence rate was 3.8%. In tumors with LVSI vs. tumors without LVSI, recurrence rate was 5.1% vs. 3.0%, respectively (p = 0.16).
The abdominal approach for radical trachelectomy is more radical in terms of parametrial and paracervical resection compared to the vaginal approach. This procedure was originally offered to patients with less favorable prognosis, e.g., tumors >2 cm and LVSI. ART can be performed by laparotomy or as minimal invasive (laparoscopic or robot-assisted) procedure. In 28 studies, 2177 patients underwent ART per laparotomy. Recurrent disease was diagnosed in 47 (3.1%) patients and 23 (1.5%) deaths were reported. Median follow-up was 47 months (range 12–120 months). In 8 studies, more detailed information on recurrences was described. Based on these studies (n
= 863 patients), recurrences were seen in 2.0% of patients with SCC, in 4.1% with AC, and in 2.7% with ASCC. These differences were not significantly different. Only one study found significantly more recurrences in ASCC [56
]. In 3 studies (n
= 367), detailed information was described on recurrences (n
= 11) and tumor size. In total, 11 patients with tumors ≥2 cm received NACT. Overall, in 3.0% of patients, a recurrence occurred of 2.3% in patients with tumors <2 cm and 3.6% with tumors ≥2 cm (p
In 10 studies, involving 534 patients, a minimal invasive approach for ART was used. Recurrence and death rates were reported in 9 studies, with a median follow-up of 44 months (range 10–98 months). In total, 15 (4.5%) recurrences occurred and 7 (1.5%) patients died of disease. In 4 studies, including 177 patients, more detailed information on recurrences (n = 14) was reported, which were significantly more often found in patients with ASCC (28.6%) than with AC (4.2%) or SCC (7.3%). Almost all recurrences (n = 12) occurred in patients with FIGO stage IB1 disease. In 2 studies with 106 patients, tumor size was specified. In 4.3% of tumors <2 cm, recurrences occurred, compared to 18.9% of tumors ≥2 cm (p = 0.03).
Some studies (n
= 24, including 3363 vs. 16,405 patients) compared oncological outcome after FSS with non-FSS (Table S8
). The disease-free survival (DFS) and overall survival (OS) or disease-specific survival (DSS) were not significantly different in both groups. In 15 studies, with 830 patients undergoing FSS and 1030 patients undergoing non-FSS, recurrences occurred in 2.4% vs. 3.3% patients, respectively. In one study, a trend was seen towards a worse DSS in FSS (conization or radical trachelectomy, n
= 125) versus non-FSS (n
= 2592) in patients with FIGO stage IB1 tumors >2 cm (82.4% vs. 90.4%, p
= 0.112) [99
In 21 of 24 studies with epithelial ovarian cancer patients, recurrence rates were reported and another 21 studies described death rates, with a median follow-up of 66 and 67 months (range 38–143), respectively.
In 15 studies, data about disease stage in relapsing patients were provided (Figure 2
). In total, 40 of 406 patients (9.9%) with stage IA/IB and 44 of 285 patients (15.4%) with stage IC had a recurrence (p
= 0.027). Stage IC is divided into three prognostic categories (IC1, IC2, and IC3) since the revised FIGO staging of ovarian cancer in 2014 [170
]. Only 8 studies used this latest FIGO staging showing that 12 of 100 patients (12%) with stage IC1/2 and 8 of 21 patients (38.1%) with stage IC3 relapsed. This difference in recurrence rate is statistically significant (p
= 0.003). The reported number of patients undergoing FSS with stage II or higher is extremely low, but 4 of the reported 10 patients (40%) with stage ≥II (stage II n
= 2, stage III n
= 2) relapsed compared to 88 of 639 patients (13.8%) with stage I (p
Death rate was 6.8% (99/1461) in stage IA/B versus 8.5% (68/800) in stage IC (p
= 0.134), 3.2% (4/120) in stage IC1/2 versus 9.5% (2/21) in stage IC3 (p
= 0.209), and 7.2% (168/2154) in stage I versus 39.2% (250/387) in stage ≥II (p
≤ 0.001) (Figure 2
In 11 studies, the grade of the carcinomas in patients who suffered from recurrences was described and complete data were reported. In 62 of 622 (9.1%) patients with grade 1 or 2 carcinoma and in 32 of 125 (25.6%) patients with grade 3 carcinoma (including clear cell histology), the cancer recurred (p ≤ 0.001). From data of 10 studies, grade 3 non-clear cell histology could be analyzed separately. These studies showed a recurrence rate of 55% (11 of 20 patients, p ≤ 0.001).
Death rate was 8.8% (219/2489) in grade 1–2 carcinomas. This is statistically significantly lower compared to grade 3 (including clear cell) carcinomas (213/711, 30.0%, p ≤ 0.001), but not to grade 3 non-clear cell carcinomas (5/26, 19.2%, p = 0.063). A high percentage (40%) of stage II or higher with worse prognosis was included in the group of grade 3 carcinomas, including clear cell histology, influencing this subgroup analysis. When limiting these calculations to stage I disease, death rate was 5.8% (115/1987) in grade 1–2 carcinomas, which is statistically significant for grade 3 (including clear cell) carcinomas (14.1%, 57/405, p ≤ 0.001) but not for grade 3 non-clear cell carcinomas (12.5%, 2/16, p = 0.239).
In total, 12 of 85 patients (14.1%) with clear cell carcinoma relapsed compared to 82 of 673 (12.2%) patients with non-clear cell carcinoma (p = 0.610). Death rates were also not statistically different (5/106, 4.7% vs. 55/933, 5.9%, respectively, p = 0.622).
Often the localization of the recurrence was not reported. In 67 patients with known location of the recurrence, an isolated ovarian tumor was found in 23 patients (34.3%) and other sites of recurrence with or without involvement of the remaining ovary in 44 patients (65.7%). Recurrences occurring in patients with initially low-risk tumors (stage IA, grade 1–2, or non-clear cell histology) were more frequently isolated (62%). There were no recurrences recorded involving the uterus.
Some studies (n
= 13) compared oncological outcome after FSS with radical, non-FSS (Table S10
). The recurrence rate of all patients who underwent FSS was 15.7% (171/1084) compared to 18.5% (266/1436) after non-FSS (p
≤ 0.001). Death rates were also lower in FSS, with 14.7% (489/3318) after FSS and 21% (1841/8774) after radical surgery (p
In 21 of 23 studies on endometrial cancer, the response rate was reported with a median follow-up of 53.8 months. Complete response (CR) occurred in 80.2% of patients (736/918). The median time to CR was 5.5 months (range 3–9 months, reported in 17/21 studies). The remaining patients (182/918, 19.8%) had persistent disease. In 13 of the 21 studies (n = 504), the response in these patients was further divided into partial response (n = 16, 3.2%), stable disease (n = 52, 10.3%), and progressive disease (n = 28, 5.6%).
In 21 studies, recurrence rates in patients with initial CR were described, which was 34.7% (297/855). The 5-year recurrence-free survival (RFS) ranged from 33 to 68% in 5 studies. Although death rate was reported in 12 studies, in only 1 study, deaths occurred (5/648, 0.8%).
The study by Chung et al. is the only study in which a new classification of endometrial cancer, based on molecular subtypes, was used [143
]. This ProMisE classification distinguishes endometrial cancer with (1) mismatch repair deficiency (MMRd), (2) DNA polymerase epsilon (POLE) mutation, (3) wild-type p53 (p53wt), and (4) abnormal p53 (p53abn) [171
]. Of the 57 patients included in this study, 9 (15.8%) had a tumor with MMRd, 2 (3.5%) had a tumor with POLE mutation, 45 (78.9%) tumors were p53wt, and 1 tumor (1.8%) had p53abn. The overall CR rate to hormonal therapy was 75.4%. Patients with MMRd tumors had a significantly lower complete or partial response rate than those with p53wt (44.4% vs. 82.2%, p
= 0.018). There was no difference in recurrence rate after achieving CR between patients with MMRd and p53wt (25.0% vs. 43.2%, p
= 0.629). The oncologic outcomes were not compared for POLE and p53abn subtypes because of the small number of patients.
Only one study, by Greenwald et al., compared hormonal therapy with primary surgery, using the Surveillance, Epidemiology, and End Results (SEER) database [149
]. Cancer-specific death rates were 3.1% (5/161) and 0.7% (46/6178), respectively (p
= 0.001). However, when comparing to a propensity score-matched cohort (1/161, 0.6%), the difference was not statistically significant (p
= 0.099). Recurrence rates were not reported.
In 2 studies, patients with grade 2 or 3 carcinoma were included and data about grade in patients who had CR were provided. Of 45 patients with grade 2–3 carcinoma, 37 (82.2%) had CR compared to 531 of 659 patients (80.6%) with grade 1 carcinoma (p = 0.787). Of these 37 patients, 15 (40.5%) had a recurrence compared to 188 of 509 patients (36.9%) with grade 1 carcinoma and CR (p = 0.661).
In 3 studies, LNG-IUD as a single treatment was used. The CR rate with LNG-IUD was 84% (63/75) versus 78.9% (435/551) with oral progestin (p = 0.309). The recurrence rate in these patients with CR was 9.5% (6/63) with LNG-IUD compared to 45.7% (217/475) with oral progestin (p ≤ 0.001). There was no difference in treatment effect between 20 or 52 mg LNG-IUD, although numbers were small. Due to heterogeneity in treatment strategies and lack of available data, no adequate subgroup analysis could be performed on type or dosage of oral progestin therapy.
3.4. Reproductive Outcome
shows the reproductive outcome after FSS per tumor type. Most studies did not report if patients had a wish to conceive and attempted pregnancy after they underwent FSS. Of the studies that did, 51% of patients with a gynecological malignancy had a pregnancy wish after FSS.
In patients with cervical cancer, 36–55% reported a pregnancy wish after FSS. A vaginal approach of FSS in cervical cancer is thought to have the best reproductive outcome, possibly due to less extensive resection of the cervix/parametrium. In patients who underwent LLETZ, conization, or ST, the pregnancy rate was 59%, with a live birth rate of 77%. After VRT, the pregnancy rate and live birth rate were 66% and 71%, respectively. The lowest pregnancy rate (45%) and live birth rate (58%) were found after ART by laparotomy, which is significantly lower than other types of FSS. In the minimal invasive approach, there was a pregnancy rate of 59%, with a live birth rate of 72%.
After the least radical FSS (LLETZ, conization, or ST), fetal loss was 15%, which is significantly lower than after VRT (21%), laparotomic ART (25%), and minimal invasive ART (24%). Furthermore, the risk of preterm delivery was also significantly lower and was 15%, 30%, 46%, and 55%, respectively.
In epithelial ovarian cancer, pregnancy wish after cancer treatment was reported in 6 studies and ranged from 25 to 65.4%, with an average of 44.2%. The live birth rate varied from 57.1 to 100%.
In 12 studies on endometrial cancer, 62.6% of patients with CR on hormonal therapy had a pregnancy wish (range 36.7–100%). Of the patients with CR, 36.9% became pregnant (reported in 17 studies).