Outcomes of ovarian transposition in gynaecological cancers; a systematic review and meta-analysis

Ovarian transposition (OT) has proven invaluable for ovarian function preservation in patients with pelvic malignancies requiring pelvic irradiation. First described in 1958 [1], it was initially performed at laparotomy until more recently when laparoscopic and robotic techniques have been described [2, 3]. Surgical approaches include intra- and retroperitoneal transposition on or lateral to the psoas muscle, the paracolic gutters, percutaneous needle transposition of the ovaries and exteriorisation to the subcutaneous fat tissue [4‐8].

It has now been established as a simple and reliable method with reduced morbidity [9]. Standard criteria for the preservation and transposition of the ovaries have been proposed [10]. In effect, the procedure is limited to a population of young premenopausal patients with early-stage, operable cervical tumours with a need for primary or adjuvant radiotherapy. It can be additionally performed for ovarian dysgerminomas, vaginal cancers but also non-gynaecological malignancies like ependymomas, Hodgkin’s disease, sarcomas and rectal carcinomas [11‐13].

Ovulation induction and oocyte retrieval can be successfully performed on transposed ovaries [14]. Studies on OT have rendered rather conflicting results, particularly those including patients who never went on to have radiotherapy. Nonetheless, retained ovaries carry a risk for symptomatic ovarian cysts and metastases to the ovaries from the primary site [15]. Yet, significant uncertainty exists regarding the efficacy of this procedure from observational studies with small sample sizes, which makes it difficult to counsel women accordingly. To ascertain the efficacy of OT, we conducted a systematic review and meta-analysis of the proportion of women who underwent OT with ovarian function preservation, symptomatic or asymptomatic ovarian cysts and metastatic ovarian malignancy.

The electronic search strategy initially yielded 396 citations. The selection process for included articles is shown in Figure 1. We retrieved 26 articles for full text examination and we further excluded a review article [18] and an article with unrelated population as it was not referring to gynaecological malignancy [19]. As a result, 24 primary studies, reporting on 892 women who underwent OT were included in this review. The main study characteristics are shown in Table 1. The quality assessment of the studies in the MINORS checklist is shown in Figure 2. All studies were observational. In brief, 7/24 (29.1%) of studies were prospective including consecutive patients in 21/24 (87.5%). There was an adequate definition of outcomes in 20/24 (83.3%) studies. No studies had a blinded assessment of the outcomes or performed a prospective calculation of sample size. The mean follow-up was longer than 12 months in 19/24 (79.1%) of studies.

In the majority of the studies, the procedure was performed in patients less than 40 years of age. Surgery was by laparotomy or laparoscopy. In 2 studies the ovaries were transposed to the subcutaneous tissue [8, 23]. The vast majority of patients (n = 828) had cervical cancer; 43 patients had vaginal cancer, 10 patients had ovarian dysgerminomas and 7 patients had pelvic sarcomas. There were 4 cases of pelvic lymphomas.

Women included in the analysis had a diagnosis of cancer and their ovaries transposed. A total of 428 women had surgery alone in the form of radical hysterectomy (RH) ± pelvic lymphadenectomy (PLND) ± upper vaginectomy ± paraaortic lymph node dissection (PALND) (Group A). 143 had post-operative brachytherapy (BR) (Group B). 321 had post-operative external beam radiotherapy (EBRT) ± BR (Group C). The follow up ranged from 2 to 126 months (Table 1).

The main aim of OT is to maintain ovarian function in premenopausal women treated with RT. It is now established as a simple and reliable method with reduced morbidity [9]. Published data vary with regards to functional outcomes such as ovarian failure, ovarian cysts and metastases to the transposed ovaries. Our systematic review of 24 studies confirms and generalizes the concept that OT is associated with a high preservation of ovarian function, an acceptable rate of ovarian cysts and a low risk of metastases in the transposed ovaries. To our knowledge, this is the first systematic review to provide an overview of the efficacy of OT on the aforementioned outcomes.

Selection of patients with early cervical and other gynaecological cancers who would benefit from OT is challenging, as difficulty arises to decide which patients would require postoperative RT prior to the surgical procedure [40]. To overcome this problem, we examined 3 treatment groups: Group A consisted of those patients who had surgery only. Patients in group B had post-operative BR. Group C consisted of patients who had primarily EBRT following surgery ± BR. OT was the fixed variable for all groups. In the absence of control groups for each treatment modality, the approach allowed for an indirect comparison amongst surgery, BR and EBRT without the risk of increasing missing data.Following a comprehensive search strategy, we employed the MINORS criteria, as indicated for quality assessment of the included observational non-comparative studies. We took into account the specific weight of the studies and after checking for heterogeneity, we employed a random-effects model to combine the data across studies to control variability. However, the majority of studies were of retrospective design. They were all observational; hence most were not designed for the specific outcomes examined with the exception of ovarian function. There was little evidence of publication bias as shown by the funnel-plot symmetry (Figure 6). The primary studies did not stratify their results by confounding factors such as age and follow up and hence, adjustment for these factors was not possible. Therefore, a certain degree of clinical heterogeneity should be expected.

In our studies, ovarian function was assessed by patient’s symptoms and serum FSH levels. Ovarian dose tolerance depends on volume irradiated, total radiation dose and the fractionation schedule in addition to patient’s age at the time of treatment. A fractionated dose exceeding 24 Gy to the ovaries can produce permanent ovarian ablation [41]. We demonstrated that ovarian function was highly preserved in those patients who had no adjuvant RT irrespective of the type of transposition or the position of transposed ovaries [25]. Only 10% of patients in that group became menopausal whilst theoretically they should be 5% or less [42]. Notably, those patients who had BR ± surgery (group B) performed as good as the surgery only group (group A). However, the strength of this inference may be reduced by the wider confidence intervals in this group due to smaller numbers of patients. While, the proportions of patients who had BR with intact uterus in group B is not known, we speculate a “timing effect” of the OT prior to irradiation as opposed to simultaneously at surgery. If OT takes place at the time of an extensive surgery, the risk for vascular compromise to the ovary from trauma or RH should be higher. Yet, the incidence of ovarian failure appears to be related to the length of follow up with 7% failing within 3 years and up to 50% within 5 years [22]. We noted that mean follow up in group B was shorter compared to group A, which might be partly responsible for the performance of the BR group. Notably, we observed that ovarian function was better preserved in those patients who have BR only without EBRT, consistent with the study by Morice et al. [32]. A plausible speculation for this is that EBRT may be damaging the vascular supply to the ovary as it loops down the pelvic brim before ascending again out of the brim to the transposed position. Therefore, if transposing an ovary prior to EBRT great care should be made to transpose the pedicle in addition to the ovary.

Further comparison between OT and non-OT cancer groups is required to assess directly the efficacy of OT for ovarian function preservation. However, considering that older women appear more susceptible to permanent ovarian damage compared to young women, it is expected that the incidence of ovarian failure following anticancer treatment ranges between 30-50% [43]. Therefore, although a direct comparison was not possible in the absence of non-OT cancer groups, we report an apparent benefit from OT for all treatment groups.

We demonstrated that the risk of ovarian carcinoma affecting the transposed ovaries is extremely low. Several risk factors for ovarian involvement have been suggested [15]. It appears that non-squamous histology carries a higher risk than the squamous one. Sutton et al. reported an incidence of 0.5% in squamous cell carcinoma compared to 1.7% in adenocarcinoma [44]. Therefore, ovarian metastasis in early cervical cancer occurs very rarely [45]. Moreover, the incidence of port site metastasis is <1% [32], which would explain our results, as the majority of transpositions were performed laparoscopically.

Risk factors for cyst development include previous surgery, extensive ovarian mobilisation albeit the mechanism is unknown and gynaecological pathology such as endometriosis or pelvic inflammatory disease [23]. In that respect, It is not surprising that the incidence of ovarian cyst formation in the transposed ovaries was common in the surgery only group (13%). In the subgroup meta-analysis, the possible advantages of subcutaneous transposition such as early detection and easier diagnosis of ovarian cysts, easy surgical access to remove ovarian cysts and facilitation of in vitro fertilization [8] were outweighed by a trend towards higher rate of ovarian cyst formation (43%) possibly due to the more extensive ovarian mobilisation at the time of subcutaneous transposition [23]. The risk for cyst formation was comparable to the BR group. As expected, only 5% of patients in the EBRT group had ovarian cysts because of the well-established association between menopausal status and ovarian cyst formation. Symptomatic cysts were identified by imaging and were treated either conservatively or surgically. Surgical treatment included needle puncture, cystectomy or oophorectomy. Conservative treatment included analgesics, hormonal or expectant management.

In conclusion, we confirm the efficacy of OT in patients undergoing radio-surgical treatment of gynaecological malignancies with high preservation of ovarian function and negligible risk of metastases to the transposed ovaries despite rather common incidence of ovarian cysts. Surgery alone and post-operative BR groups performed best for the above outcomes. As quality of care remains an important issue in cancer care, careful expansion of patient selection could identify those premenopausal patients who would really benefit from this rather underutilised procedure.