Fertility-sparing surgery for women with stage I cervical cancer of 4 cm or larger: a systematic review

1. Search strategy

The preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines were used for this study [10]. A search was performed up to September 30, 2020 by three authors (C.G., S.C.M., S.G.) independently within several database (MEDLINE, Embase, Cochrane Library, Google Scholar) to ensure all relevant studies evaluating the impact of FSS for stage FIGO 2018 IB3, FIGO 1994–1997 stage IB1–IB2 or FIGO 2009 stage IB1–IB2 cervical cancer on oncological and reproductive outcomes.

The process of evidence acquisition combined the following MeSH terms: “cervical cancer,” “cervix,” “conservative surgery,” “conservative treatment,” “fertility-sparing surgery,” “trachelectomy,” “conization,” “cone resection,” and “fertility.” Article abstracts, full text of articles and cross-referenced studies identified from retrieved articles were screened for pertinent information. All duplicate records were removed.

2. Selection of studies and methodologic quality assessment

Key criteria for inclusion were: 1) original studies published in English, in peer-reviewed journals studies; 2) stage I cervical cancer of 4 cm or greater in size at the time of initial diagnosis; 3) desire of the patients to preserve their fertility and refusal of radical treatment to be stated; and 4) completed FSS (i.e., trachelectomy, conization).

Exclusion criteria were: 1) editorials, review articles and conference abstracts; 2) studies with incomplete or absent data on obstetric outcomes; and 3) studies not reporting surgical and fertility outcomes of the cohort of patients with cervical cancer of 4 cm or larger separately from the study population.

The selected studies were comprehensively examined, and relevant data extracted for each paper and inputted to the spreadsheet. The information selected included: journal, authors, year of publication, age, race, histotype, grading, tumor size at diagnosis, lymphovascular space invasion (LVSI), neoadjuvant chemotherapy (NACT), tumor dimension after NACT and response rate, type of FSS (abdominal radical trachelectomy [ART], vaginal radical trachelectomy [VRT], vaginal simple trachelectomy [VST], and conization), surgical approach (surgical access route: abdominal versus vaginal), type of nodal assessment (sentinel lymph node, pelvic and/or paraaortic lymphadenectomy) and approach used (laparotomy, laparoscopy, robotic), number of resected lymph nodes, number of positive lymph nodes, follow-up schedule, number of pregnancies (including miscarriages) for each patient, type of conception (spontaneous conception vs assisted reproduction), type of delivery (term vs preterm), recurrence rate, disease-free survival (DFS), and overall survival (OS). The three authors (C.G., S.C.M., S.G.) carried out data extraction and quality assessment from all the retrieved studies based on full-text articles. Discrepancies between the investigators were resolved by consensus.

3. Outcomes

The primary outcome of interest was the impact of FSS on oncological and fertility outcomes. In particular, for each eligible study cohort, the following outcomes were considered:

4. Statistical analysis

Survival curves were estimated for each group, considered separately, using the Kaplan-Meier method and compared statistically using the log rank test. As studies differed for different follow-up time frame, in order to reduce the risk of survival bias, a 5-years landmark analysis technique was applied by manually excluding patients with follow-up durations below or above the landmark's threshold.

1. Primary studies included in the literature review

A search of the MEDLINE (PubMed) database resulted in 766 relevant articles and further search in the Embase, Cochrane Library, Google Scholar databases yielded no additional articles. No additional eligible studies were retrieved by hand-searching bibliographies. Fifteen studies fulfilled the inclusion criteria for this systematic review, involving a total 48 patients (Fig. S1).

2. Study characteristics

The characteristics of included studies are detailed in Table 1. Among the 15 studies included, 9 studies (60%) were prospective [9, 11, 14, 15, 16, 17, 18, 19, 20], 2 (13.3%) were retrospective [8, 21], and 4 (26.7%) were case reports [22, 23, 24, 25]. No randomized controlled trials were found. Eight studies (53.3%) were monocentric [8, 16, 17, 18, 22, 23, 24, 25] and seven (46.7%) were multicentric [9, 11, 14, 15, 19, 20, 21]. Three studies (20%) were published between 2005 and 2010 [14, 22, 23], while 12 studies (80%) were published between 2011 and 2019 [8, 9, 11, 15, 16, 17, 18, 19, 20, 21, 24, 25]. The definition of inclusion/exclusion criteria for FSS were available for 11 studies (73.3%) [8, 9, 11, 14, 15, 16, 17, 18, 19, 20, 21] (Table S1). Of note, one study reported as inclusion criteria for VRT tumors with stage IA1 with LVSI, IA2, and IB1 with no more than 2 cm; however, it reported also the case of one patient with 45-mm stage IB2 cervical cancer treated with VRT after NACT [18]. Four studies (26.7%) did not report exclusion criteria as they were case reports [22, 23, 24, 25]. The most reported exclusion criteria were nodal metastases (10/11; 90.9%) followed by positive margins (5/11; 45.5%), distant metastasis (3/11; 27.3%), age >40 years (2/11; 18.2%), and aggressive histotypes (2/11; 18.2%).

Table 1
Characteristics of included studies

• Click for larger image
• Click for full table
• Download as Excel file

Details regarding preoperative instrumental and laboratorial assessment were available for 13 studies (86.7%) [8, 9, 11, 15, 16, 17, 18, 19, 20, 21, 23, 24, 25] (Table S2). Pre-treatment evaluation consisted of magnetic resonance imaging (MRI) in the vast majority of the studies (12/13; 92.3%) followed by computed tomography (CT) (6/13; 46.2%), positron emission tomography (PET) (6/13; 46.2%), and colposcopy (5/13; 38.5%). Six (40%) out of 15 studies described the specific purpose of imaging. MRI [2, 15, 18, 21, 25] was used to determine tumor size and parametrial involvement, while CT [15, 17, 21, 25] and PET [17, 21] scans were carried out to evaluate the presence of lymph node or distant metastases. Of the 11 authors who included patients undergoing NACT, 3 did not report evaluation methods after NACT [17, 18, 22], while 2 others evaluated the response to NACT through imaging but did not specify the technique [8, 25]. The criteria for defining the response to NACT were not unanimous throughout the included studies: 4 studies [8, 11, 17, 22] reported the pathological response to NACT, while 6 [9, 16, 19, 23, 24, 25] the radiological one. The most adopted assessment tools were MRI and colposcopy (5/6, 83.3%), followed by ultrasound (3/6, 50%).

The follow-up schedule was tailored individually for 34 patients (70.8%) with a weighted mean±standard deviation of 56.15±38.75 months. The remaining 14 patients (29.2%) were extracted from 3 cohorts who reported a median follow up of 90 months [20], 90 months [15], and 32 months [14], respectively.

3. Patient characteristics and surgical data

The characteristics of included patients are summarized in Table 2. For each variable, missing data are reported in Table 2.

Table 2
Characteristics of patients selected (n=48)

• Click for larger image
• Click for full table
• Download as Excel file

Precise age was provided for 18 patients (37.5%) with a weighted mean of 28.61 years (median age: 28 years; 25th–75th percentile: 25.75–31.25 years). The remaining 30 patients (62.5%) were extracted from cohorts who reported median age ranging between 28 and 32 years [8, 20] and mean age ranging from 30.5 and 32.5 years [14, 15, 17, 21], respectively.

Data about histology were reported for 39 patients (81.3%): 25 (64.1%) presented with squamous cell carcinoma, 10 (25.6%) with adenocarcinoma, 3 (7.7%) with glassy cell carcinoma, and 1 (2.6%) with adenosquamous carcinoma. Tumor grading was available for 16 patients (33.3%): 8 patients (50%) had grade 2 tumor, 5 (31.2%) grade 3, and 3 (18.8%) grade 1.

Thirty-three patients underwent NACT and 14 did not. The maximum tumor diameter at diagnosis was reported in exact millimeters for 34 patients (70.8%), whereas it was only known to be ≥4 cm for the remaining 14 patients (29.2%). The weighted mean tumor maximum diameter at diagnosis was 44.59±5.71 mm (median: 42 mm; 25th–75th percentile: 40–50 mm) (NACT group [n=33]: 44.42±5.7 mm [median: 42 mm; 25th–75th percentile: 42–47.5 mm]; non-NACT group [n=1]: 50 mm). The LVSI was specified for 10 patients (20.8%), being positive in 5 patients (50.0%) and negative in 5 patients (50.0%).

Among the 33 patients who underwent NACT, data about post-NACT tumor size were reported for 32 patients. The weighted mean post-NACT tumor size was 7.25±10.87 mm (median: 0 mm; 25th–75th percentile: 0–14.25 mm). Seventeen patients (53.1%) had a complete response (11 pathological [8, 11, 17, 22] and 6 radiological [16, 19, 25]), 12 (37.5%) a partial response (4 pathological [8] and 8 radiological [9, 16, 19, 23, 24]) and 3 (9.4%) a stable disease (pathological [8]).

All 48 patients underwent trachelectomy as fertility-sparing surgery. The most adopted surgical route was open abdominal (31 out of 48 procedures; 64.6%), followed by vaginal (17/48; 35.4%). Among the 17 patients who underwent vaginal trachelectomy, one was robotic-assisted (5.9%), 5 (29.4%) were laparoscopic-assisted, while 11 (64.7%) were exclusively vaginal. Furthermore, among patients who underwent vaginal trachelectomy, 14 (82.4%) were radical (1 robotic-assisted and 13 purely) and 3 (17.6%) were simple.

Data about lymphadenectomy were available for 47 patients. Forty (85.1%) patients underwent pelvic lymphadenectomy, 7 (14.9%) women had both pelvic and para-aortic lymphadenectomy. The most used surgical approach for lymphadenectomy was laparoscopy (60.4%), followed by open abdominal surgery (37.5%) and robotic surgery (2.1%). Lymphadenectomy was performed in all patients: before FSS in 8 patients (16.7%) and during FSS in 40 patients (83.3%). Of note, the timing of lymphadenectomy among patients receiving NACT differed between the studies: lymphadenectomy was the first-step during FSS in 27 (81.8%) patients [8, 11, 17, 18, 19, 23, 24, 25], while it was performed before NACT in 6 (18.2%) patients [9, 16, 22]. The exact number of removed lymph nodes was available for 30 patients (62.5%) with a weighted mean of 22.4±9.58 (median: 19; 25th–75th percentile: 17–25.5). The remaining 18 patients (37.5%) were extracted from three cohorts that reported a median of 24 [15, 21] and a mean of 32.2 lymph nodes being removed [14].

4. Oncological outcomes

Three women (6.3%) experienced a recurrence and one of them (2.1%) died of disease. Therefore, the occurrence of a relapse was fatal in 33% of cases. The median DFS was not reached and the 5-year DFS rate was 92.4%. The 5-year OS rate was 97.6%.

There was no statistically significant difference between patients who received NACT before FSS and those who underwent upfront FSS in terms of 5-year DFS (91% vs. 92.9%, respectively; p=0.946) and 5-year OS (100% vs. 92.9%, respectively; p=0.173).

Although lacking the statistical significance, a trend towards worse 5-year DFS rate was reported for adenocarcinoma (71.4%) and glassy cell (66.7%) histologies compared with squamous cell (log-rank test, p=0.070) (Fig. 1A), for grade 3 (50%) compared with grade 1–2 (100%) (log-rank test, p=0.060) (Fig. 1B), and for tumors ≥5 cm (74.1%) compared with those <5 cm (94.4%) (log-rank test, p=0.131).

Fig. 1
Survival outcomes, Kaplan-Meier plots for patients with early-stage cervical cancer ≥4 cm treated with FSS. (A) DFS according to histotype; log-rank (Mentel-Cox), p=0.070. (B) DFS according to histologic grading; log-rank (Mentel-Cox), p=0.060. (C) DFS according to risk group; log-rank (Mentel-Cox), p=0.024.
DFS, disease-free survival; FSS, fertility-sparing surgery.

• Click for larger image
• Download as PowerPoint slide

Given the strict interdependency between these three factors, patients were classified into 2 subgroups: 1) “high-risk” if they had at least one of the following factors: adenocarcinoma/glassy cell histology, G3, or diameter ≥5 cm; 2) “low-risk”, if none of these risk factors was present. A significantly lower 5-year DFS was reported for high-risk patients compared with low-risk (74.7% vs. 100%; log-rank test, p=0.024) (Fig. 1C).

Moreover, concerning the three patients who reported a recurrence, the first patient had a diagnosis of 5-cm glassy cell carcinoma and underwent ART plus pelvic lymphadenectomy. A pelvic side-wall recurrence occurred 5 months after primary treatment and, due to the failure of chemo-radiotherapy, a pelvic exenteration was performed. The patient died of disease 16 months after FSS [20]. The second recurrence occurred in a 37-year-old patient affected by G3 adenocarcinoma of the cervix. The patient underwent laparoscopic staging plus lymphadenectomy and 2 cycles of NACT with a partial response (62.5%). Twenty months after VRT, the patient had a pelvic recurrence successfully treated with posterior exenteration and chemoradiation [16]. The third patient had a 5-cm tumor (adenocarcinoma) and underwent NACT with a complete response. A recurrence was diagnosed 45 months after ART and the patient underwent hysterectomy with lymphadenectomy. In July 2018, the patient was alive without evidence of disease [8].

5. Fertility outcomes

Data about fertility were available for 12 patients (25%): 5 attempted to conceive (41.7%) [9, 17, 24], while 7 did not (58.3%) [17, 19, 23]. Most relevant data about those patients are summarized in Table 3. Of the 5 patients who attempted to conceive, 4 conceived (pregnancy rate: 80%) with a total of 6 pregnancies (patient #2 and #4 had 2 pregnancies including an ectopic pregnancy) and gave birth to five children in total (live birth rate: 83.3%). Four pregnancies were carried at term (80%) [17], whereas one ended with a preterm labor (premature delivery rate: 20%) [24]. In all five cases, a caesarean section (CS) was performed. Concerning the type of radical trachelectomy, 3 of these patients (3/4; 75%) had been treated with open ART [17, 24], while 1 with VRT (1/4; 25%) [9]. All 4 patients underwent NACT before trachelectomy. Unfortunately, none of the studies reported data about the following surgical treatment after delivery, whether patients underwent standard treatment. Given the small number of patients who attempted to conceive, no statistical analysis was performed.

Table 3
Obstetrical outcomes and characteristics of pregnant patients after FSS

• Click for larger image
• Click for full table
• Download as Excel file

Briefly, the first patient managed to conceive through in vitro fertilization-embryo transfer (IVF-ET) after five attempts with artificial inseminations [24]. The pregnancy was complicated by threatened abortion and a cervical cerclage was performed at 13 weeks of gestation. Absolute bed rest was prescribed from around 16 weeks. Rupture of the membrane occurred at 27 weeks and 8 days later an emergent CS was performed. After 5 years, the baby was growing healthily. The second patient, who had a 4-cm tumor and underwent ART after NACT, had 2 pregnancies, both ended with a term delivery through CS [17]. The third patient within the ART group underwent CS at term, but data regarding the type of conception and the course of pregnancy were missing [17]. Finally, the only patient in the VRT group conceived spontaneously after a first ectopic pregnancy managed with methotrexate [9, 26]. There were no significant consequences related to the previous trachelectomy (i.e., no cervical shortening, vaginal bleeding, or threatened preterm labor); the patient developed gestational diabetes treated with insulin. An elective CS was performed at 38 weeks of gestation. The new-born was healthy and weighted 3,500 g and the infant development was normal.