Aiming to create a transplantable artificial ovary, our goal was to optimize our previous follicle isolation protocol in order to increase the number and survival of isolated human preantral follicles [
9,
10]. While the previous method does yield a high number of isolated preantral follicles, we always observed extruded oocytes [
9,
10] and undigested fragments of tissue. This was probably due to the prolonged and fixed duration of enzymatic digestion (75 min). For this reason, we hypothesized that a successful isolation protocol should take into account two important points: first, the high degree of heterogeneity in follicle density, not only between women and age groups [
13,
25], but even between fragments of ovarian tissue from the same patient [
14]; and second, the variations in extracellular matrix composition between women of different age [
15,
16] and health status [
15,
18] and even between different areas of the same ovarian tissue sample [
16,
27], which influences success rates of the isolation procedure. To meet these requirements, the follicle isolation protocol should be tailored to each patient; instead of a fixed period for enzymatic digestion, as currently reported in the literature [
7‐
11], this step would be halted as soon as all the fragments are totally digested. In addition, enzymatic digestion would be divided into several time intervals, in order to recover already isolated follicles, which would reduce the proportion of damaged follicles due to prolonged digestion. Therefore, with a view to maximizing the yield and quality of isolated ovarian follicles for each patient, we developed a tailored follicle isolation protocol that addresses the two limitations encountered with our earlier method [
9,
10], namely long and fixed enzyme exposure time. Comparison of the two isolation protocols showed that with our modifications, based on fractioning enzymatic digestion over three different durations, we could obtain a greater number of isolated follicles than with the previously applied protocol. Since 35% of isolated follicles were already identified after the initial 30 min of enzymatic digestion in the modified protocol, we can confirm our hypothesis that the long period of enzyme exposure with the previous method has a negative impact on follicles. Although dispase present in high concentrations in Liberase DH does not cleave laminin [
28], which is one of the main components of the basement membrane of human primordial follicles [
29], it cleaves collagen IV [
28], another essential constituent of the basement membrane of these follicles [
29]. Hence, one can assume that follicles isolated during the first few minutes of enzymatic digestion could have their basement membrane progressively damaged over the course of 75 min of exposure to the enzyme, which would probably explain the presence of extruded oocytes in our previous studies [
9,
10]. As also demonstrated by histological analysis, the number of primordial and primary follicles was significantly higher after fibrin encapsulation in the modified protocol (92.3%) than in the earlier one (73.4%), which more closely reflects the physiological distribution of follicles inside ovarian cortex [
19,
30]. Our findings therefore confirm that discontinuous enzymatic digestion positively affects recovery rates of isolated follicles by protecting their structure from unnecessary exposure to Liberase DH. Interestingly, the percentage of secondary follicles was greater with the previous method than the modified version (26% versus 9%), but similar to an earlier report [
12]. Unlike in the previous protocol, a filtration step was also added to the new protocol in order to replace follicle centrifugation and separate follicles from ovarian cell suspensions. We could hypothesize that the different percentage of secondary follicles found between the two protocols may be ascribed to some secondary follicles remaining trapped inside the filter and thus getting lost. With the modified isolation protocol, we even recovered a significant number of viable isolated follicles after 90 min of enzymatic incubation, which shows that the ovarian tissue fragments that remained undigested in our previous protocol are actually an important but as yet unexploited source of primordial follicles. Since every follicle counts towards increasing the likelihood of pregnancy in patients and the possibility of repeating this strategy, discarding such precious material may have a negative impact on artificial ovary outcomes. Another important point to bear in mind is that the artificial ovary will be mostly applied to prepubertal cancer patients. The ovaries of these young girls have a softer cortex than that of adult women, since the ovaries accumulate collagen fibers with age [
31,
32] making the cortex stiffer. It would therefore be logical to assume that follicle isolation should be considerably faster in young patients than in adult women [
13], so it is likely that the 75-min digestion time in our previous protocol would damage a significant number of these follicles. In order to confirm the effectiveness of the modified isolation protocol for our artificial ovary prototype, one fibrin clot containing isolated follicles and SCs was grafted for a short period of time. Interestingly, with the modified protocol, the superior follicle recovery rate after grafting (35%) was better than in a previous study (20–23%), where follicles were isolated using the earlier protocol and grafted inside a fibrin-based matrix [
21] for one week. Moreover, in the present study, a higher percentage of primordial follicles (72% versus 15%) but a similar proportion of primary follicles (28% versus 26%) were found on day 7 of transplantation compared to our previous findings [
21]. Given the superior preservation of the primordial follicle pool with the modified protocol, it could be interesting to further investigate whether this protocol plays a role in post-grafting activation of follicles.