Ovarian tissue grafting: Lessons learnt from our experience with 55 grafts

Abstract Purpose Uncertainties remain regarding the clinical efficacy of ovarian tissue cryopreservation and grafting. We report a retrospective analysis of reproductive outcomes and lessons learnt following 55 ovarian tissue transplant procedures at our center from 2006 to 2019. Methods We analyzed variables related to graft success such as tissue volume, follicular density, total follicular volume, and age on the duration of graft function. Results Follicular density and total follicular volume correlate positively with duration of graft function. All clinical pregnancies in our cohort occurred in women who were aged 35 or less at the time of ovarian tissue cryopreservation. Conclusion Graft success, as determined by eventual pregnancy and the longevity of graft function, may be impacted by factors including age at cryopreservation, follicular density, and total follicular volume.


| INTRODUC TI ON
A major side effect of cancer treatment in young women, such as chemotherapy and radiation therapy, is the loss of ovarian function and fertility, which has a significant impact on quality of life. 1 Given the relatively high incidence of cancer in reproductive age women and improvements in survival, an increasing number of women are presenting for discussion of ways to preserve their fertility prior to undergoing gonadotoxic treatment. 2 A number of strategies have been developed in recent years, including medical therapies to suppress ovarian function, improvements in cryopreservation of oocytes or embryos, and ovarian tissue cryopreservation (OTC). 3 Ovarian tissue cryopreservation can be performed at any time without delaying the cancer treatment and is the only available option to preserve fertility in children and premenarchal adolescent girls. 4 Harvesting part or a whole ovary allows the cryopreservation of potentially thousands of oocytes in their immature form, whereas oocyte freezing will result in storage of only a finite number of mature oocytes with a limited number of potential opportunities. 5 The procedure may not only benefit those with oncological diseases, but also those with benign diseases that require conditioning regimens for bone marrow transplant (BMT) such as aplastic anemia or thalassemia or those with autoimmune conditions and genetic disorders associated with the development of premature ovarian insufficiency (POI). 6,7 Since the first study that described OTC and grafting in 2000 8 and first live birth reported in 2004, 9 over 140 babies have been born worldwide. 3,10 However, uncertainties remain regarding its clinical efficacy, optimal grafting site, tissue volumes, and factors which may predict success. 5 As thousands of women have had ovarian tissue frozen, increasing numbers are returning for ovarian tissue grafting, making exploration of these issues critical. 11 We report a retrospective analysis of reproductive outcomes following 55 ovarian tissue transplant procedures at our center from 2006 to 2019, and analyze variables related to graft success such as tissue volumes and follicular density.

| MATERIAL S AND ME THODS
Initially, tissue freezing was approved under Human Research and Ethics provided by Royal Women's Hospital (RWH). Following extensive validation of procedures in xenografted models, 12 Clinical Review Board of RWH gave permission for clinical transplantation pilot study. The oncological (81%) and medical (19%) indications for ovarian tissue freezing are shown in Table 1. Non-oncological indications included imminent POI, severe endometriosis, benign ovarian cysts, and autoimmune conditions such as Wegener's granulomatosis and systemic lupus erythematosis, which may have the potential to compromise future fertility. Oncological indications included Hodgkin's and non-Hodgkin's lymphoma, granulosa cell ovarian tumor, breast, uterine and cervical cancer, Ewing's sarcoma, and soft tissue sarcoma. Ovarian tissue harvest was performed laparoscopically, with the exception of those requiring laparotomy for treatment of their medical condition. The tissue was harvested prior to gonadotoxic treatment in all but 10 patients (Table 1).

| Statistical analysis
Descriptive statistics are used throughout to describe various outcomes. Statistical analysis was restricted to determining correlation coefficients and was performed by calculating Pearson's correlation coefficient and producing scatter plots with lines of best fit to visually demonstrate the strength of association. STATA version 16.1 (StataCorp) was used.

| Collection and cryopreservation of ovarian tissue
In three patients, unilateral oophorectomy was performed due to high risk of POI from impending BMT and radiotherapy treatment.
Partial oophorectomy was performed in the remainder, avoiding the use of electrocoagulation. No surgical complications such as infection, bleeding, or organ damage were observed in this group of patients. Histology and immunohistochemistry, where appropriate, were performed on a sample of retrieved tissue to exclude malignant infiltration and to estimate follicle density. Follicular density was estimated as the average number primordial follicles counted in at least two pieces of cortex per millimeter of ovarian cortex. There was no evidence of malignancy in any samples of the tested tissue.
The ovarian tissue was trimmed to 1 mm depth and subsequently sliced (approximately 1 mm wide by 3-4 mm long) to aid cryoprotectant (propanediol) penetration and cryopreserved according to our previously published protocols. [12][13][14] Larger pieces of cortex (5 × 5 × 1 mm) were also frozen in some cases. Vials containing multiple slices were stored in nitrogen vapor, and for these patients, the duration of storage was between 1.5 and 16.7 years (Table 1).
For tissue cryopreserved at other centers ( Table 2), less detailed information regarding the volume of tissue or size of pieces or cryopreservation procedure was available. Slow freezing was used for all cases regardless of center where freezing was performed; however, no specific details were given apart from cryoprotectants used at the other center: propanediol (patient 3,5,15,18,19,29), dimethyl sulfoxide (patient 7,12, 32,37), and ethylene glycol (patient 38).

| Transplantation of thawed ovarian tissue
Grafting of stored ovarian tissue was considered for patients who presented with POI requesting fertility assistance. The age of the patients at the time of first grafting ranged between 25.1 and 47.2 years (median age 36). Two patients were particularly interested in restoration of endocrine function in addition to or instead of a specific desire for fertility. Menopausal status was confirmed by cessation of menstruation and persistently elevated FSH over 40 IU/L in all the patients. AMH levels were below 0.1 pmol/L in all patients.
The patients' medical fitness was considered and in all cases suitability to proceed with fertility and pregnancy was confirmed by the treating medical specialist.
In situations where there was concern about uterine functionality for pregnancy following radiation, pelvic ultrasound, MRI and the adequacy of endometrial response to artificial hormonal stimulation were utilized to assess uterine functionality for pregnancy.

TA B L E 1 (Continued)
Prior to grafting, a small amount of frozen cortex and a was thawed and subjected to detailed histopathology and immunohistochemistry where appropriate. No malignancies were detected in any of the samples examined. The ovarian tissue was rapidly thawed and rehydrated as previously described. 15

| Graft sites
In patients with tissue frozen at our center, ovarian slices were threaded onto a 6.0 vicryl suture horizontally to maximize area of contact with peritoneum to facilitate neovascularization (generally 10-15 slices per suture) with total tissue volumes ranging 60-575 mm 3 ( Table 3). The total number of slices available determined the amount of tissue put back in the first procedure (generally half of total frozen), to leave some tissues available for future use. Ovarian and past treatment, such as radiotherapy. Volume of tissue grafted could not be estimated for patients 43 and 44 (Table 2) due to tissue having been minced into small fragments these were introduced to pockets using a pipelle.

| Monitoring
The management plan for patients following graft surgery, involved ultrasound and endocrine assessment including follicle stimulating hormone (FSH), luteinizing hormone (LH), progesterone (P4) and estradiol (E2) at 4 weeks and monthly thereafter. When endocrine activity resumed, as indicated by a drop in FSH (<20 U/ml) and rise in estradiol levels (>200 pg/L), ovarian function was monitored with ultrasound and endocrine assessment tracked for one to three cycles, after which time an IVF cycle was commenced if appropriate for those patients requesting fertility.

| Restoration of ovarian activity
Of the 55 grafts (Table 1), 40 grafts belonged to the 29 patients who underwent cryopreservation at our center (Group 1; Table 3) and 14 grafts were from the 11 patients who had tissue cryopreserved in other units (Group 2; Table 2). Ovarian activity as demonstrated by a rise in serum E2 and fall in FSH, occurred in 36/40 patients (90%) after a mean of 4.5 months (range 1-9 months). In Group 1, activity was observed in 29/29 patients (100%; Table 3). For those patients in Group 2, activity resumed in 9/11 patients (82%; Table 2).

| Duration of endocrine function
Long-term follow-up was conducted for up to 13 years (2006-2019), with serial endocrine assessments until cessation of endocrine function, as defined by consistently elevated FSH levels, patients declining further surveillance or treatment, or lost to follow-up.
The duration of ovarian activity in those whose function ceased was between 0-65 and 0-24 months, in groups 1 and 2, respectively.
The longest functioning graft was 65 months in a patient with treated non-Hodgkin's lymphoma, who underwent unilateral oophorectomy and cryopreservation following cyclophosphamide-containing chemotherapy (patient 1).

| Stimulated cycles and embryo transfers
Twenty-eight patients underwent repeated stimulation cycles (Tables 2 and 3

| Monitoring for disease relapse
Pathology testing for malignant infiltration was performed at the time of cryopreservation and again prior to grafting. The potential risk of malignant infiltration was explained and written information was provided to patients.
In patient 10, with a history of granulosa cell tumor, 16

| Factors predicting duration of function
The effect of tissue volume, follicle numbers, total follicular number, and age at cryopreservation on duration of ovarian function were analyzed, as shown below.

Tissue volume
There was a weakly positive correlation between the duration of function and tissue volume (Pearson's r = .08) but this was not statistically significant (P = .73). This analysis excludes those whose function did not resume. Patients without tissue volume and duration of function recorded at time of harvest were also excluded. For patients with multiple grafts, the graft with the longest corresponding functional duration was used for analysis ( Figure 1).

Follicle density
Follicle density positively correlated with the duration of function (Pearson's r = .67), which was highly statistically significant (P = .0008). Patients without follicle density and number of folli- with multiple grafts, the graft with the longest corresponding functional duration was used for analysis ( Figure 3).

Age at cryopreservation
There was a statistically significant inverse correlation between the age at harvest and duration of ovarian function (Pearson's r = −.46) which is statistically significant (P = .03). For those patients who had more than one graft procedure, the graft with the longest duration of activity was used for analysis. These data exclude those whose function did not resume and those with ongoing function ( Figure 4).

| Grafting sites
The site from which oocytes were recovered which subsequently formed embryos that resulted in a pregnancy are reported in Table 4.
All were single embryo transfers except patients 9 and 10. Patient 10 had two embryos transferred, both from oocytes from the abdominal graft, which resulted in twin birth previously reported. 16

| Effect of prior chemotherapy
Ten patients underwent chemotherapy treatment prior to ovarian tissue harvest and cryopreservation. Parameters potentially affected by chemotherapy include the follicular density, time to resumption of function and duration. The numbers are not statistically significant, and the sample size is too small to demonstrate correlation between chemotherapy and ovarian function.

| D ISCUSS I ON
Graft success, as determined by eventual pregnancy and the longevity of graft function, may be impacted by factors including age at cryopreservation, baseline ovarian reserve, techniques of tissue preparation and transplantation, history of cancer treatment, freezing-thawing protocols, number of cortical sections grafted, transplantation techniques and graft sites, and degree of tissue ischemia after transplantation. 18 Some studies have suggested that follicular density or ovarian tissue volume at the time of grafting could influence ovarian transplantation outcome. 3,19 Our study has shown a positive correlation between follicle density and total follicle number on duration of graft function. This could be due to the fact that patient numbers are relatively small and cancer survivors are a heterogeneous group of patients with multiple factors influencing outcome. 19 The empty follicle rate in our study is 32.2%. Studies have shown that follicles developing in the transplanted ovarian tissue are more often without oocyte and the pregnancy rate from IVF is relatively low compared to standard IVF. 20 An empty follicle rate as high as 23%-35% had been reported in some studies after ovarian function at 2-4 months after second grafting, which persisted for 9-84 months. This is thought to be due to a combination of angiogenic or hormonal factors. 28 After the initial follicular depletion in the first graft, the second graft is subject to a more stable endocrine environment and less follicular loss. The rescue mechanism is not fully understood but may be due to diffusion of paracrine substances secreted from the grafted ovarian tissue to prevent apoptosis of granulosa cells in the endogenous ovary and activate residual follicles. 30 For this reason, like some other units, we do not use all available tissue in the first grafting procedure. 3,31,32 There is no gold standard for grafting site, as it is still unclear which of the grating locations are superior. 32 51 which has one of the highest conception rates after ovarian tissue grafting.
Grafts from tissue prepared and frozen at other clinics showed reduced function and longevity. It is difficult to comment on why this is the case. For those who may not have local access to a unit with specialist ovarian tissue processing expertise, there is now experience worldwide with tissue transportation with a cooling device, for up to 22 hours, prior to freezing. 52 There is an increasing number of successful pregnancies and live births from transported tissue. 9,52 This method has the potential to improve women's access to fertility preserving options, as well as standardizing outcomes between specialized and smaller centers. Transplantation of cryopreserved ovarian tissue is recognized as an established technique for fertility preservation and despite numerous publications from major centers, uncertainties still remain regarding its clinical efficacy, optimal grafting sites, factors which may predict success, as well as safety. We address these factors by reporting our experience with 55 ovarian tissue transplant procedures performed at a single center. Future direction in this field should include the optimization of transplantation reproductive outcomes, longer duration of endocrine function and improved safety profile.

D I SCLOS U R E S
Genia Rozen, Stephanie Sii, Franca Agresta, Debra Gook, Alex Polyakov and Catharyn Stern declare that they have no conflict of interest.

E TH I C A L A PPROVA L
The protocol for the research project has been approved by a suitably constituted Ethics Committee.

H U M A N/A N I M A L R I G HTS
The article does not contain any studies with human and animal subjects performed by the any of the authors.