{"paper_id":"d1fdd4dc-0ea9-415c-b13c-5f6622b6f96b","body_text":"RESEARCH\nJournal of Robotic Surgery          (2026) 20:576 \nhttps://doi.org/10.1007/s11701-026-03415-4\nYunjeong Park and Eunah Shin contributed equally to this work and \nare designated as co-first authors.\n \r Joo Hyun Park\nbeanpearl@yuhs.ac\n1 Institute of Women’s Life Medical Science, Yonsei \nUniversity College of Medicine, Seoul  \n03722, Korea, Republic of\n2 Department of Obstetrics and Gynecology, Severance \nHospital, Yonsei University College of Medicine, 50 -1 \nYonsei-ro, Seodaemun-gu, Seoul 03722, Korea, Republic of\n3 Department of Pathology, Yongin Severance Hospital, Yonsei \nUniversity College of Medicine, Seoul  \n16995, Korea, Republic of\n4 Department of Obstetrics and Gynecology, Gangnam \nSeverance Hospital, Yonsei University College of Medicine, \nSeoul 06229, Korea, Republic of\n5 Department of Obstetrics and Gynecology Yongin Severance \nHospital, Yonsei University College of Medicine,  \nSeoul 16995, Korea, Republic of\nAbstract\nOvarian endometrioma cystectomy may compromise ovarian reserve through inadvertent excision of ovarian cortex. We \ncompared inadvertent cortical removal between robotic-assisted and conventional laparoscopic cystectomy using digital \npathology. We retrospectively analyzed 81 patients (40 laparoscopic, 41 robotic) who underwent single-surgeon cystec -\ntomy (January 2020–December 2025) with digitized hematoxylin and eosin–stained slides available. Ninety-eight ovary/\nside specimens were classified as follicle-containing cortex, cortex without follicles, or fibrosis, and excised cortical area \n(mm²) was quantified. The primary analysis used log-linear regression adjusted for cyst length and width with patient-\nclustered robust standard errors. Tissue-type distribution did not differ by approach ( P = 0.611). Excised cortical area was \nsmaller with robotics (median 34.6 mm², interquartile range 16.9–82.3) than laparoscopy (median 65.4 mm², interquartile \nrange 39.5–81.6; P = 0.011). In the adjusted model, robotics was associated with a smaller excised cortical area (robot-to-\nlaparoscopy ratio 0.55, 95% confidence interval 0.32–0.93; P = 0.029). Follicle counts and antral follicle presence among \nfollicle-containing specimens were comparable. Robotic-assisted cystectomy was associated with less inadvertent excision \nof ovarian cortex after accounting for cyst dimensions, while follicle-based specimen metrics did not differ.\nKeywords Ovarian endometrioma · Laparoscopic surgery · Robotic surgery · Digital pathology · Ovarian cortex · \nOvarian reserve\nReceived: 1 March 2026 / Accepted: 3 April 2026\n© The Author(s) 2026\nRobotic versus laparoscopic enucleation of ovarian endometriotic \ncysts with pathological analysis of inadvertent follicular loss\nYunjeong Park1,2 · Eunah Shin3 · Jimin Bae1,4 · SiHyun Cho1,4 · Young Sik Choi1,2 · Joo Hyun Park1,5\nAbbreviations\nAFC  Antral follicle count\nAMH  Anti-Müllerian hormone\nASRM  American Society for Reproductive Medicine\nBMI  Body mass index\nCI  Confidence interval\nH&E  Hematoxylin and eosin\nIQR  Interquartile range\nIPW  Inverse probability weighting\nOR  Odds ratio\nBackground\nEndometriotic cysts of the ovaries are one of the most com-\nmon presentations of endometriosis, occurring in up to 40% \nof women with the condition and up to 50% of women with \ninfertility [ 1]. While endometriotic cysts frequently cause \nclinical symptoms including dysmenorrhea, pelvic pain, \ninfertility, and potential increased risk of malignant trans -\nformation, the surgical intervention of these ovarian lesions \n\n1 3\n  576  Page 2 of 10\nJournal of Robotic Surgery          (2026) 20:576 \npresents a crucial clinical dilemma: the need to adequately \nremove the ectopic endometrial tissue within the ovarian \ntissue while losing less ovarian reserve at the same time \n[2–4]. There is compelling evidence demonstrating that sur-\ngical excision (cyst enucleation) of endometriotic cysts, no \nmatter how meticulous, is paradoxically linked to signifi -\ncant impairment of ovarian reserve [ 5]. This consequence \nmay compromise future fertility and potentially accelerate \nreproductive aging [6].\nThe mechanisms underlying this iatrogenic tissue dam -\nage during ovarian surgery are regarded as being multifac -\ntorial. During enucleation of ovarian cysts and adhesiolysis, \nunintended removal of healthy ovarian tissue represents \none critical pathway to follicular or ovarian volume loss, \nas endometriotic cysts are infiltrative in nature and the \ncyst pseudo-capsules often lack distinct clean histological \nboundaries from the surrounding tissues [ 7]. Histopatho -\nlogical examination of specimens from ovarian cyst enucle-\nation consistently indicates that recognizable ovarian tissue \nexists adjacent to the cyst wall in over 90%, containing fol-\nlicles of all stages. In addition to mechanical stripping, vas-\ncular compromise and thermal injury during hemostasis, as \nwell as postoperative inflammation/edema, may contribute \nto acute and subacute follicle loss [8, 9].\nThese observations have led some clinicians to defer \nendometrioma surgery before assisted reproduction unless \ncyst size/location or refractory pain warrants intervention. \nHowever, the possibility of spontaneous pregnancy resto -\nration and effective pain amelioration after surgical inter -\nvention argues compellingly to find an optimized surgical \nmodality which minimizes iatrogenic ovarian damage [ 2]. \nAdvancement of technology in surgical instrumentation has \nled to robotic-assisted platforms, the most widely used being \nthe da Vinci Surgical System. Robotic surgical systems \noffer ergonomic advantages over conventional laparoscopy, \nwhere articulation of the robotic tools provide more freedom \nof motion, three-dimensional binocular visualization with \nmagnification, enhanced precision accompanied by tremor \nfiltration and less surgeon fatigue [10, 11]. However, there is \nlittle evidence in the literature supporting the actual benefits \nof utilizing the robotic armament in preventing unwanted \nfollicle loss during ovarian cyst enucleation [12].\nDirect microscopic quantification of ovarian cortex and \nfollicles inadvertently attached to the excised cyst wall pro-\nvides granular insight into technique-related tissue loss [ 7, \n12]. Using standardized analysis of digitized H&E slides, \nwe aimed to determine whether robotic assistance—via \nenhanced visualization and instrument control—translates \ninto reduced inadvertent excision of ovarian cortex and folli-\ncle-bearing tissue compared with conventional laparoscopy.\nMethods\nStudy population\nThe study was performed at the Department of Obstetrics \nand Gynecology, Yongin Severance Hospital, Yonsei Uni -\nversity Medical Center, encompassing a total of 81 patients \n(40 laparoscopic and 41 robotic) who received either da \nVinci robotic (Xi or SP) or conventional laparoscopic \nenucleation (cystectomy) of endometriotic cysts between \nJanuary 2020 and December 2025. Only cases performed \nby a single gynecologic surgeon for both robotic and lap -\naroscopic surgery were enrolled to avoid inter-surgeon \ntechnical variability. The operating surgeon had extensive \nexperience in minimally invasive gynecologic surgery, \nincluding more than 1,600 da Vinci Xi/SP cases and more \nthan 6,000 laparoscopic cases over 13 years at a university \ncenter. A retrospective analysis of the pathologic slides was \nperformed to compare the amount (area) of attached folli -\ncle-containing healthy ovarian cortical tissue surrounding \nthe pseudo-capsule of the enucleated endometriotic cysts. \nPatients were excluded if they had concomitant ipsilateral \novarian pathology other than endometrioma, suspected or \nconfirmed malignancy/borderline tumor, pregnancy at the \ntime of surgery, prior ovarian surgery, hormonal suppres -\nsion (oral contraceptives or other agents such as progestins/\ndienogest or GnRH agonist/antagonist) within 3 months \nbefore surgery, or were younger than 20 years or older than \n38 years.\nEthical consideration\nThe study was approved by the Institutional Review Board \nof Yongin Severance Hospital and conducted in accordance \nwith the Declaration of Helsinki (IRB No. 9-2026-0013). \nThe requirement for informed consent was waived due to \nthe retrospective design.\nSurgical procedures\nRobotic surgery was performed using either the da Vinci Xi \nmultiport platform or the da Vinci SP single-port platform. \nIn Xi cases, a transumbilical multichannel port was placed \nat the umbilicus for camera and assistant access, and two \nadditional robotic working ports were placed bilaterally \n8 cm lateral and 2 cm caudal to the umbilicus. In SP cases, \na single transumbilical multichannel port was inserted \nthrough the umbilicus, through which the robotic camera, \nrobotic instruments, and assistant access were introduced. \nIn conventional Anthrex laparoscopy, a three-port technique \nwas used, consisting of a 5-mm 30° umbilical camera port \nand two 5-mm working ports placed at the McBurney and \n\n1 3\nPage 3 of 10   576 \nJournal of Robotic Surgery          (2026) 20:576 \ncontralateral McBurney points. Standard laparoscopic bipo-\nlar forceps and monopolar scissors or a monopolar hook \nwere used for dissection, with grasping forceps or Allis \nforceps for traction; a laparoscopic needle holder was used \nwhen hemostatic suturing was required. Specimens were \nretrieved in an endoscopic bag through the umbilical port/\nincision.\nRobotic cases used bipolar forceps, monopolar curved \nscissors or a monopolar hook, a wristed needle driver, and \na ProGrasp or Cadiere forceps; laparoscopic cases used \nstandard bipolar forceps, monopolar scissors or hook, and \ngrasping forceps, including Allis forceps when stronger \ntraction was required. Across all approaches, cyst enucle -\nation began with the smallest feasible cortical incision at the \nthinnest identifiable ovarian cortex after adhesiolysis. Trac-\ntion-countertraction was used to identify the least vascular \ncleavage plane, and the dissection plane was re-routed when \nthe initial plane appeared suboptimal. Hydrodissection with \nsaline injected through an intraperitoneal needle was used \nwhen feasible, although it was often limited in densely \ninfiltrative endometriomas. Energy use was minimized \nthroughout dissection. After enucleation, hemostasis was \nfirst attempted with gentle compression using hemostatic \ngauze. Bipolar coagulation was restricted to pinpoint appli-\ncation for focal bleeding away from the ovarian hilum or \ninfundibulopelvic ligament, whereas broader oozing areas \nwere managed with tension-free 3 − 0 polyglactin (Vicryl) \nsutures. The hemostatic strategy did not materially differ \nbetween the Xi and SP platforms.\nHistologic analysis\nThe specimens were transferred immediately to the pathol -\nogy department for gross examination and fixation. After \ntissue fixation, hematoxylin and eosin-stained tissue slides \nwere digitized using the Philips Fast Scanner (Philips \nElectronics, The Netherlands) and analyzed. Whole-slide \nimaging enabled full-slide review on screen with digital \nannotation of ovarian cortical boundaries and automated \ncalculation of annotated area, allowing more standardized \nand reproducible quantification than manual visual estima -\ntion by conventional light microscopy [ 13]. The pathologi-\ncal review and initial annotations were performed by a single \npathologist specialized in gynecologic and breast pathology, \nwith secondary quality-control review by another patholo -\ngist in a randomly selected subset of cases. A representative \nslide containing the highest tissue area with the least frag -\nmentation from each case of endometriotic cyst enucleation \nwas screened first for the presence or absence of ovarian \ncortical tissue attached to the cystic wall. One representative \nslide per case, the slide containing the largest cortical tis -\nsue area with the least fragmentation, was selected because \nthe number of paraffin blocks and slides varied across cases \naccording to cyst size and specimen fragmentation; this \napproach was used to standardize comparisons across cases. \nThe specimens were classified into the following three cat -\negories: (1) endometriotic cyst with attached ovarian corti -\ncal tissue containing follicles; (2) endometriotic cyst with \nattached ovarian cortical tissue but without follicles; and \n(3) endometriotic cyst specimen with predominant fibrosis \n(fibrotic tissue) without identifiable follicles. For those with \novarian cortical tissue attached, the total area of attached \novarian cortical tissue was obtained by delineating the corti-\ncal tissue areas on the digitized slide and summing the areas. \nThe ovarian cortical tissue showing follicles embedded \nwithin was analyzed further for the total number of follicles \nobserved, ranging from primordial to antral. Follicle density \nwas determined by obtaining the ratio of the number of fol-\nlicles per ovarian cortical tissue area (mm²). Cortical tissue \ndelineation was performed at x50 or ×100 magnification, \nfollicle counting at ×200, and follicle classification at x200 \nor ×400. For reproducibility assessment, a random quality-\ncontrol subset of approximately one in five cases was inde-\npendently reviewed by a second pathologist from the same \ninstitution.\nSample size\nThe sample size calculation was based on the preliminary \ndata presented by Sinha et al. (2024) [ 12], using a two-\nsided independent samples t-test with α = 0.05, β = 0.20 \n(80% power), and assuming a moderate standardized effect \nsize (Cohen’s d = 0.68) (approximately corresponding to \na difference of 3 follicles between groups), resulting in a \nminimum sample size of 35 per group. To account for an \napproximately 10% anticipated specimen inadequacy and \nplanned subgroup analyses by bilateral versus unilateral \ndisease, we aimed to enroll at least 40 patients per group. \nWith 40 patients per group, the planned sample size pro -\nvides approximately 85% power for d = 0.68 under these \nassumptions.\nStatistical analysis\nThe analysis combined two cohorts: conventional lapa -\nroscopy (lapa) and robot-assisted surgery (robot). Patient-\nlevel variables (e.g., age, BMI, preoperative AMH, ASRM \nstage, cyst length/width, and bilaterality, and exploratory \nclinical outcomes) were summarized at the patient level. \nDigital pathology outcomes (tissue type and follicle met -\nrics) were analyzed at the ovary/side specimen level (left \nor right). Because some patients contributed bilateral speci-\nmens, regression models for specimen-level outcomes used \ncluster-robust standard errors clustered at the patient level.\n\n1 3\n  576  Page 4 of 10\nJournal of Robotic Surgery          (2026) 20:576 \nContinuous variables were summarized as median (inter-\nquartile range [IQR]) and compared between groups using \nthe Mann–Whitney U test. Categorical variables were sum-\nmarized as n (%) and compared using the chi-square test or \nFisher’s exact test, as appropriate. Exploratory patient-level \nclinical and fertility-related outcomes included estimated \nblood loss, perioperative hemoglobin change (preoperative \nminus postoperative hemoglobin), perioperative complica -\ntion, transfusion, 30-day readmission, reoperation, docu -\nmented recurrence, chart-documented pregnancy intention, \nand documented pregnancy among patients with pregnancy \nintention when available. All tests were two-sided; P < 0.05 \nwas considered statistically significant and values < 0.001 \nare reported as P < 0.001.\nTissue type categories were defined as described in the \nHistologic analysis section (type 1: follicle-containing cor -\ntex; type 2: cortex without follicles; type 3: predominant \nfibrosis). For follicle-related outcomes, blank antral follicle \ncounts were treated as zero when tissue type was known. \nBlank total follicle counts were treated as zero for tissue \ntypes 2 or 3, and treated as missing for tissue type 1 to avoid \nconflating missing entry with a true zero. Cystic follicle \ncounts were available only in the robotic cohort and were \nsummarized descriptively. Because tissue type 2 (with -\nout follicles) may reflect both surgical technique and low \nbaseline ovarian reserve, multivariable models prioritized \nadjustment for age and preoperative AMH.\nThe primary outcome was ovarian cortical area (mm²) \nmeasured in the excised specimen. Because cortical area \nwas right-skewed, it was log-transformed. The prespeci -\nfied primary model was a log-linear regression: log(cortical \narea) = β0 + β1·(robot vs. lapa) + β2·log(cyst length) + \nβ3·log(cyst width). Exponentiated coefficients exp(β1) are \nreported as the robot-to-laparoscopy ratio (ratio < 1 indi -\ncates smaller excised cortical area in the robot cohort). Pre-\nspecified additional adjustments included age, bilaterality, \nand ASRM stage (modeled as categorical). Because excised \ncortical area can scale with cyst size, cyst length and width \nwere included as covariates in the primary model (log-trans-\nformed) to adjust for lesion size.\nSecondary outcomes included: (i) the probability that a \nspecimen contained follicle-bearing tissue (tissue type 1 vs. \nothers); and among tissue type 1 specimens, (ii) the presence \nof any antral follicle (binary), and (iii) follicle and antral \nfollicle counts. Binary outcomes were modeled using logis-\ntic regression with patient-clustered standard errors. Count \noutcomes were modeled using Poisson regression with an \noffset of log(cortical area) to estimate density (counts per \nunit area). Age and preoperative AMH were included as \ncovariates to partially account for baseline ovarian function.\nSensitivity analyses included [ 1] restricting analyses to \nASRM stage 4 cases; and [ 2] inverse probability weight -\ning (IPW) to address differential availability of cortical area \nmeasurements between cohorts. For IPW, the probability \nof having cortical area recorded was modeled using logis -\ntic regression including surgery type, age, ASRM stage, \nlog(cyst length), log(cyst width), and bilaterality. Stabilized \nweights were applied to the primary cortical-area model, \nand weights were additionally trimmed at the 1st and 99th \npercentiles in a robustness check.\nStatistical analyses and figure generation were performed \nusing R version 4.4.2 (R Foundation for Statistical Comput-\ning, Vienna, Austria), with complete-case inclusion for each \nmodel.\nResults\nStudy population and data completeness\nThe analytic cohort included 81 patients (40 lapa and 41 \nrobot) contributing 98 ovary/side specimens (44 lapa and 54 \nrobot). Among the robotic cohort, 34/41 (82.9%) underwent \nsurgery with the da Vinci SP platform and 7/41 (17.1%) \nwith the da Vinci Xi platform. Tissue type classification was \navailable for all included specimens by design. Cortical area \nwas recorded in 37/44 (84.1%) lapa specimens and 49/54 \n(90.7%) robot specimens.\nIn the evaluated cohort, age, BMI, and preoperative \nAMH did not differ materially between groups (Table 1). \nThe robot cohort had more advanced disease by ASRM \nstage (median 4.0 vs. 3.0; P = 0.008) and larger cyst length \n(median 4.65 vs. 4.00 cm; P = 0.031).\nPathology tissue type and follicle metrics\nRepresentative histologic examples illustrating cortical \ntissue delineation, tissue-type classification, and follicle \nVariable Lapa (median [Q1,Q3]) (n = 4 0 ) Robot (median [Q1,Q3]) (n = 4 1 ) P value\nAge, years 29.50 [27.00, 36.25] 30.00 [25.00, 33.00] 0.457\nBMI, kg/m² 20.91 [19.61, 22.47] 21.39 [19.51, 24.43] 0.529\nPreoperative AMH 2.96 [1.47, 4.97] 3.54 [2.00, 4.93] 0.435\nASRM stage 3.00 [3.00, 4.00] 4.00 [4.00, 4.00] 0.008\nCyst length, cm 4.00 [3.00, 5.00] 4.65 [3.50, 6.12] 0.031\nCyst width, cm 2.60 [2.00, 4.00] 3.35 [2.00, 4.05] 0.425\nBilateral disease 9/40 (22.5%) 16/41 (39.0%) 0.149\nTable 1 Baseline characteristics \n\n1 3\nPage 5 of 10   576 \nJournal of Robotic Surgery          (2026) 20:576 \nsubtype identification are shown in Figs. 1, 2 and 3. Tissue \ntype distribution did not differ between surgical approaches \n(chi-square P = 0.611; Table 2). Follicle-bearing tissue (type \n1) was observed in 86.4% of lapa specimens and 79.6% of \nrobot specimens. Among type 1 specimens, follicle count \nand antral follicle metrics were similar across groups \n(Table 2). In adjusted logistic regression, the odds of type 1 \ntissue were not significantly different for robot vs. lapa after \nadjustment for age (OR 0.61, 95% CI 0.20–1.80; P = 0.370). \nIn the age + AMH model, higher preoperative AMH was \nassociated with increased odds of type 1 tissue (Supplemen-\ntary Information: Table S1). In this quality-control subset, \nno material discrepancies affecting cortical-area delineation \nor follicle counting were identified on secondary review.\nPrimary outcome: ovarian cortical area\nThe excised cortical area was smaller in the robot cohort \n(median 34.55 mm², IQR 16.94–82.30) compared with lapa \n(median 65.42 mm², IQR 39.48–81.59; P = 0.011; Table 2 \nand Fig. S1). In the primary log-linear model adjusting for \ncyst length and width (both log-transformed), robot-assisted \nsurgery was associated with a lower cortical area (robot-to-\nlapa ratio 0.55, 95% CI 0.32–0.93; P = 0.029). The associa-\ntion was consistent after additional adjustment for age (ratio \n0.55; P = 0.040) and after further adjustment for bilaterality \nand ASRM stage (ratio 0.52; P = 0.027) (Table 3 and Figs. \nS2–3).\nSensitivity analyses\nIn ASRM stage 4 cases only, the association remained in \nthe same direction but was borderline significant (ratio \n0.53, 95% CI 0.29–0.99; P = 0.054; Table 3). IPW analy -\nses addressing differential recording of cortical area \nyielded similar estimates (IPW primary ratio 0.55, 95% CI \n0.32–0.93; P = 0.030), and results were robust to trimming \nextreme weights (ratio 0.54, 95% CI 0.31–0.92; P = 0.026; \nTable 3).\nAdditional perioperative and exploratory follow-up out -\ncomes are summarized in Supplementary Table S2. Esti -\nmated blood loss was higher in the robotic group, although \nconcomitant myomectomy was also more frequent in robotic \ncases. Other short-term safety outcomes and postoperative \nhormonal treatment variables were broadly similar between \ngroups. Recurrence and pregnancy-related data were based \non chart documentation during nonuniform follow-up and \nshould be interpreted descriptively.\nDiscussion\nIn this study, we applied digital pathology–based quantifica-\ntion of ovarian tissue inadvertently excised during ovarian \nendometrioma cystectomy and compared robotic-assisted \nversus conventional laparoscopic approaches. The principal \nfinding was that the cortical area present in the excised cyst \nwall specimen was smaller in the robotic cohort, and this \nassociation persisted after adjustment for cyst size and other \nFig. 1 Ovarian cortical tissue delin-\neated. The area of ovarian cortical \ntissue attached on the outer surface \nof the endometriotic cyst was care-\nfully delineated (x50)\n \n\n1 3\n  576  Page 6 of 10\nJournal of Robotic Surgery          (2026) 20:576 \nclinically relevant covariates. In contrast, the distribution \nof tissue-type categories (follicle-containing cortex vs. fol -\nlicle-absent cortex vs. fibrosis) and follicle subtype counts \n(including antral follicles) did not differ significantly by sur-\ngical approach. Together, these results suggest that robotic \nassistance may reduce inadvertent removal of ovarian cor -\ntex, while the follicular content of the excised specimen—a \nmetric influenced by both the surgical dissection plane and \nbaseline ovarian reserve—may be less sensitive to surgical \napproach in a real-world cohort.\nOvarian reserve is commonly assessed in clinical prac -\ntice using serum anti-Müllerian hormone (AMH) and ultra-\nsound-based antral follicle count (AFC), which reflect the \nremaining follicular pool and correlate with ovarian respon-\nsiveness [ 14]. However, interpreting these biomarkers in \nthe setting of ovarian surgery requires nuance. Serum AMH \nrepresents the combined contribution of both ovaries; there-\nfore, following unilateral surgery, systemic AMH is not side-\nspecific and may incompletely capture ipsilateral tissue loss, \nparticularly when contralateral function is preserved [ 15, \n16]. Moreover, AFC measurement is susceptible to cycle-to-\ncycle fluctuation and operator dependence. Standardization \nstatements have emphasized uniform acquisition and report-\ning, yet variability remains [ 17]. These limitations provide \nstrong rationale for direct specimen-based quantification of \ninadvertent ovarian tissue loss, as performed in the present \nstudy using digital pathology.\nPrior studies have shown that ovarian endometrioma \ncystectomy can be associated with a substantial reduction \nin ovarian responsiveness. In a paired IVF-cycle analy -\nsis comparing operated versus contralateral intact ova -\nries after unilateral endometrioma excision, Somigliana \net al. reported a mean 53% reduction in dominant follicle \ndevelopment in the previously operated ovary [ 18]. Ragni \net al. further suggested that this impairment reflects pre -\ndominantly quantitative rather than qualitative injury, with \nreduced numbers of follicles/oocytes/embryos but similar \nfertilization and embryo quality measures [ 19]. Systematic \nreviews and meta-analyses consistently demonstrate post -\noperative AMH decline after stripping cystectomy [ 6, 20]. \nFig. 2 Ovarian tissue type. A Ovarian cortical tissue containing fol -\nlicles (mainly primordial, black arrow) was classified as tissue type \n1 (x50). The luminal surface of the endometriotic cyst is noted (red \narrow). B Ovarian cortical tissue without follicles (black arrow) was \nclassified as tissue type 2 (x20). The luminal surface of the endometri-\notic cyst is noted (red arrow). C The cortical tissue shows no identifi-\nable follicle within the ovarian cortical tissue in tissue type 2 (x400). \nD The endometriotic cyst wall is sole composed of hyalinized fibrotic \ntissue with inflammatory cells and vessels without any ovarian tissue \nattached (black arrow, x20). The luminal surface of the cyst is noted \n(red arrow)\n \n\n1 3\nPage 7 of 10   576 \nJournal of Robotic Surgery          (2026) 20:576 \nImportantly, ovarian reserve compromise appears greater \nin bilateral disease in comparative syntheses [ 20], although \nbilaterality does not necessarily translate into worse IVF/\nICSI outcomes in all post-cystectomy series [ 21]. When \nAMH and AFC were evaluated concurrently in the same \nwomen, AMH declines were consistently detected across \npostoperative windows whereas AFC changes were smaller \nand less consistent [ 22]. Together, these data highlight that \nexcised cortical area is best interpreted as a surrogate histo-\nlogic endpoint of iatrogenic ovarian tissue loss, which may \ncomplement but does not directly substitute for functional \noutcomes such as AMH, AFC, pregnancy, or live birth.\nRobotic-assisted surgery may offer technical fea -\ntures—stable three-dimensional visualization, articulated \nFig. 3 Ovarian follicle types identified within tissue type 1. A Pri -\nmordial and primary follicles (x400). The black arrow shows a pri -\nmordial follicle with a single layer of squamous granulosa cells and \nthe red arrow shows a primary follicle with a single layer of cuboidal \ngranulosa cells. B Intermediate primordial follicle (x600) with s single \nlayer of squamous and cuboidal cells. Late primary follicle (x400) sur-\nrounded by multiple layers of cuboidal cells. D Small antral follicle \n(x200). E Antral follicle (x100) with more than 50% of the follicle \nshowing antrum\n \n\n1 3\n  576  Page 8 of 10\nJournal of Robotic Surgery          (2026) 20:576 \ninstruments, and improved ergonomics—that could facili -\ntate a more precise dissection plane and reduce inadver -\ntent inclusion of normal cortex. Evidence supporting this \nhypothesis has begun to emerge. Sinha et al., using artificial \nintelligence–assisted whole-slide imaging to quantify tis -\nsue loss in excised specimens, reported that robotic assis -\ntance reduced ovarian tissue loss, particularly in bilateral \ndisease and with increasing cyst size [ 12]. Our findings are \nconcordant with respect to area-based tissue loss, but we \ndid not observe consistent between-group differences in fol-\nlicle counts or antral follicle presence. This pattern supports \nthe interpretation that excised cortical area may be a more \nstable surrogate of inadvertent ovarian tissue removal than \nfollicle counts alone in retrospective cohorts, where follicle-\nbased endpoints are sparse and biologically heterogeneous.\nThe absence of between-group differences in follicle \nsubtype counts and tissue-type categories warrants careful \ninterpretation. Follicles observed within excised cyst wall \nspecimens are not purely a marker of surgical technique; \nthey are also influenced by baseline ovarian reserve, patient \nage, AMH, endometrioma-related cortical distortion, and \nthe thickness of the cyst wall–cortex interface. In clinical \nterms, “no follicles detected” may reflect successful cleav -\nage along the correct plane, but may also occur when the \npatient’s follicle density is intrinsically low. Accordingly, \nfollicle-related endpoints should be interpreted as “mixed” \nsignals reflecting both operative performance and ovarian \nbiology, and adjusted analyses incorporating age and preop-\nerative AMH are important to reduce confounding. In line \nwith this interpretation, preoperative AMH was indepen -\ndently associated with the presence of follicle-containing \ntissue (type 1) in adjusted analysis.\nWe view cortical area within the excised specimen as \na surrogate histologic marker of iatrogenic ovarian tissue \nremoval and a complementary, pathology-grounded indi -\ncator of tissue-sparing surgery, rather than a direct func -\ntional measure of postoperative ovarian reserve or fertility \npotential. Unlike serum AMH—which reflects systemic \ngranulosa-cell activity from both ovaries—specimen cor -\ntical area directly captures the physical quantity of cor -\ntex removed from the operated side. However, whether a \nsmaller excised cortical area translates into improved AMH, \nAFC, pregnancy, or live-birth outcomes remains uncertain \nand requires prospective validation. Thermal damage from \nhemostasis and devascularization may impair remaining \ntissue without increasing the amount of cortex present in \nthe specimen. Randomized trials have evaluated hemostatic \nstrategies (e.g., bipolar coagulation vs. suturing) with vari -\nable effects on postoperative ovarian reserve markers [ 8, \n23], highlighting that tissue-sparing dissection and tissue-\npreserving hemostasis likely operate in parallel.\nThe clinical implications of minimizing iatrogenic ovarian \ndamage are substantial. Surgery for ovarian endometrioma \ncan improve pain and may reduce recurrence, yet concerns \nabout diminished ovarian reserve have prompted ongoing \ndebate regarding surgical timing and the role of proceeding \ndirectly to assisted reproduction in selected patients. Con -\ntemporary guidelines emphasize individualized decision-\nmaking, balancing symptom control and fertility goals, and \nrecognize the need to minimize ovarian injury when surgery \nis undertaken [ 3]. This emphasis on fertility-aware, tissue-\nsparing surgery is consistent with broader trends in benign \ngynecologic surgery away from unnecessarily radical treat-\nment [24], while long-term patient-centered outcomes such \nTable 2 Pathology findings (specimens with tissue type)\nOutcome Lapa\n(n = 4 4 )\nRobot\n(n = 5 4 )\nP value\nTissue type 1 (with follicle) 38 (86.4%) 43 (79.6%) 0.611\nTissue type 2 (without follicle) 4 (9.1%) 6 (11.1%)\nTissue type 3 (fibrosis) 2 (4.5%) 5 (9.3%)\nCortical area (mm²), median [IQR] (area recorded) 65.42 [39.48, 81.59] (n = 3 7 ) 34.55 [16.94, 82.30] (n = 4 9 ) 0.011\nCortical area in tissue type 1 only (mm²), median [IQR] 67.08 [50.73, 81.62] (n = 3 5 ) 39.60 [26.43, 96.02] (n = 4 2 ) 0.062\nFollicle count in tissue type 1, median [IQR] 16.5 [5.0, 77.5] (n = 3 4 ) 23.5 [11.0, 38.5] (n = 4 2 ) 0.810\nAntral follicle count in tissue type 1 (0 included), median [IQR] 0.00 [0.00, 0.75] (n = 3 8 ) 0.00 [0.00, 1.00] (n = 4 3 ) 0.952\nAntral follicle present in tissue type 1, n/N (%) 10/38 (26.3%) 12/43 (27.9%) 1.000\nModel Robot/Lapa ratio 95% CI P value\nUnadjusted (log area ~ surgery) 0.529 0.316–0.887 0.016\nPrimary: +log(cyst L) + log(cyst W) 0.546 0.321–0.929 0.029\nPrimary + age 0.550 0.313–0.964 0.040\nPrimary + age + bilateral + ASRM stage 0.518 0.292–0.916 0.027\nSensitivity: stage 4 only (primary) 0.534 0.287–0.995 0.054\nSensitivity: IPW (primary) 0.545 0.318–0.934 0.030\nSensitivity: IPW trimmed (1st–99th pct, primary) 0.535 0.312–0.919 0.026\nTable 3 Primary outcome regres-\nsion (exp(beta)=Robot/Lapa \nratio)\n \n\n1 3\nPage 9 of 10   576 \nJournal of Robotic Surgery          (2026) 20:576 \nas pain relief and quality of life remain essential alongside \npathologic endpoints [25].\nThis study has strengths. First, it leverages digital \npathology to quantify tissue endpoints in an objective and \nreproducible manner, addressing limitations of clinical bio-\nmarkers and ultrasound measures. Second, specimen-level \nanalyses accounted for within-patient correlation, and sen -\nsitivity analyses adjusted for cyst dimensions and disease \nseverity.\nNevertheless, several limitations should be acknowl -\nedged. First, the retrospective, nonrandomized design limits \ncausal inference. Baseline differences between groups—\nnotably in ASRM stage and cyst size—together with the \nincomplete availability of cortical area measurements, \nparticularly in the laparoscopic group, may have intro -\nduced selection bias. Although multivariable adjustment \nand inverse probability weighting were performed, residual \nconfounding is still likely. Learning-curve bias is another \npotential concern in any retrospective single-surgeon com -\nparison. In the present study, however, laparoscopic and \nrobotic cases were performed during the same study period \nby a surgeon with extensive experience in both minimally \ninvasive platforms, making a major platform-level learning-\ncurve effect less likely. Even so, subtle temporal changes \nin case selection, operative judgment, or procedure-specific \ntechnique cannot be fully excluded. The predominance of \nSP cases within the robotic cohort likely reflects nonrandom \nplatform selection for adnexal surgery in reproductive-age \nwomen, in whom a transumbilical single-incision approach \nmay be attractive because it minimizes visible abdominal \nscarring. Accordingly, the present findings should be inter -\npreted as comparing a predominantly SP robotic cohort with \nconventional laparoscopy, rather than as a platform-specific \ncomparison between the Xi and SP systems. In addition, \nother unmeasured factors, including specific dissection and \nhemostasis techniques and specimen handling, may also \nhave contributed to the observed differences. Although the \ninitial annotations were performed by a single specialized \npathologist, a random quality-control subset was cross-\nchecked by a second pathologist without material dis -\ncrepancies; nevertheless, formal full-cohort interobserver \nagreement statistics were not available. In addition, to stan-\ndardize comparisons across cases with variable block num-\nbers and tissue fragmentation, cortical area was quantified \nfrom one representative slide per case; therefore, the total \namount of inadvertently excised cortex across all blocks \nmay have been underestimated. Finally, because systemic \novarian reserve and fertility outcomes were not uniformly \navailable longitudinally, the extent to which reduced excised \ncortical area translates into improved postoperative ovarian \nfunction requires prospective validation. Accordingly, these \nfindings should be interpreted as pathology-based evidence \nof reduced inadvertent cortical excision using a surrogate \nhistologic endpoint, rather than direct proof of improved \nlong-term ovarian reserve or fertility outcomes. Future \nstudies should integrate specimen-based metrics with stan -\ndardized ovarian reserve testing, symptom/quality-of-life \noutcomes, fertility endpoints, and emerging biomarker-\nbased phenotyping approaches [26].\nConclusion\nIn conclusion, robotic-assisted cystectomy was associated \nwith reduced inadvertent excision of ovarian cortex as mea-\nsured by excised cortical area, while follicle-based speci -\nmen metrics did not differ consistently between approaches. \nThese findings support the potential for robotic assistance \nto reduce iatrogenic ovarian tissue loss, particularly rel -\nevant for patients with fertility goals, bilateral disease, or \nlarger cysts. Prospective studies incorporating standardized \nreserve assessment and longer-term reproductive outcomes \nare warranted to confirm clinical benefit.\nSupplementary Information  The online version contains \nsupplementary material available at  h t t  p s : /  / d o  i . o  r g / 1 0 . 1 0 0 7 / s 1 1 7 0 1 - 0 \n2 6 - 0 3 4 1 5 - 4     .  \nAcknowledgements Not applicable.\nAuthor contributions Y .P. and E.S. contributed equally to this work. \nY .P. wrote the manuscript and performed the statistical analysis. E.S. \nperformed all pathological analyses and wrote parts of the manuscript, \nincluding preparation of all figures. J.B. performed the retrospective \nchart review and organized the data. S.C. and Y .S.C. advised on the \nstudy design and statistical methodology. J.H.P. conceived and de -\nsigned the study, reviewed the patient data, and wrote the manuscript. \nAll authors read and approved the final manuscript.\nFunding The authors declare that no funds, grants, or other support \nwere received during the preparation of this manuscript.\nData availability The datasets used and/or analyzed during the cur -\nrent study are available from the corresponding author on reasonable \nrequest.\nDeclarations\nCompeting interests The authors declare no competing interests.\nEthics Approval This study was approved by the Institutional Review \nBoard of Yongin Severance Hospital (IRB No. 9-2026-0013) and was \nconducted in accordance with the Declaration of Helsinki.\nConsent to Participate  The requirement for informed consent was \nwaived by the IRB due to the retrospective nature of the study and the \nuse of de-identified data/specimens.\nConsent to Publish Not applicable. This manuscript does not contain \nany individual person’s identifiable data in any form (including indi -\n\n1 3\n  576  Page 10 of 10\nJournal of Robotic Surgery          (2026) 20:576 \nvidual details, images, or videos).\nClinical Trial Number Not applicable.\nOpen Access   This article is licensed under a Creative Commons \nAttribution-NonCommercial-NoDerivatives 4.0 International License, \nwhich permits any non-commercial use, sharing, distribution and \nreproduction in any medium or format, as long as you give appropri -\nate credit to the original author(s) and the source, provide a link to the \nCreative Commons licence, and indicate if you modified the licensed \nmaterial. You do not have permission under this licence to share \nadapted material derived from this article or parts of it. The images or \nother third party material in this article are included in the article’s Cre-\native Commons licence, unless indicated otherwise in a credit line to \nthe material. If material is not included in the article’s Creative Com -\nmons licence and your intended use is not permitted by statutory regu-\nlation or exceeds the permitted use, you will need to obtain permission \ndirectly from the copyright holder. To view a copy of this licence, visit \nhttp:   //creativecommo ns. o rg/lice ns es /b y -nc-nd/4.0/.\nReferences\n1. Gałczyński K, Jóźwik M, Lewkowicz D, Semczuk-Sikora A, \nSemczuk A (2019) Ovarian endometrioma - a possible finding in \nadolescent girls and young women: a mini-review. J Ovarian Res \n12(1):104\n2. Muzii L, Galati G, Mattei G, Chinè A, Perniola G, Di Donato V et \nal (2023) Expectant, medical, and surgical management of ovar -\nian endometriomas. J Clin Med 12(5):1858\n3. Becker CM, Bokor A, Heikinheimo O, Horne A, Jansen F, Kiesel \nL et al (2022) ESHRE guideline: endometriosis†. Hum Reprod \nOpen 2022:2\n4. Barnard ME, Farland LV , Yan B, Wang J, Trabert B, Doherty JA \net al (2024) Endometriosis Typology and Ovarian Cancer Risk. \nJAMA 332(6):482–489\n5. Younis JS, Shapso N, Fleming R, Ben-Shlomo I, Izhaki I (2019) \nImpact of unilateral versus bilateral ovarian endometriotic cystec-\ntomy on ovarian reserve: a systematic review and meta-analysis. \nHum Reprod Update 25(3):375–391\n6. Raffi F, Metwally M, Amer S (2012) The impact of excision of \novarian endometrioma on ovarian reserve: a systematic review \nand meta-analysis. J Clin Endocrinol Metab 97(9):3146–3154\n7. de Koning R, Blikkendaal MD, de Sousa Lopes SMC, van der \nMeeren LE, Cheng H, Jansen FW et al (2024) Histological analy-\nsis of (antral) follicle density in ovarian cortex tissue attached to \nstripped endometriomas. J Assist Reprod Genet 41(4):1067–1076\n8. Deckers P, Ribeiro SC, Simões RDS, Miyahara C, Baracat EC \n(2018) Systematic review and meta-analysis of the effect of bipo-\nlar electrocoagulation during laparoscopic ovarian endometrioma \nstripping on ovarian reserve. Int J Gynaecol Obstet 140(1):11–17\n9. Song T, Kim WY , Lee KW, Kim KH (2015) Effect on ovarian \nreserve of hemostasis by bipolar coagulation versus suture during \nlaparoendoscopic single-site cystectomy for ovarian endometrio-\nmas. J Minim Invasive Gynecol 22(3):415–420\n10. Park Y , Song A, Jee J, Bae N, Oh S, Shin JH et al (2024) Changes \nin anti-Müllerian hormone values for ovarian reserve after mini -\nmally invasive benign ovarian cystectomy: comparison of the Da \nVinci robotic systems (Xi and SP) and the laparoscopic system. \nSci Rep 14(1):9099\n11. Truong M, Kim JH, Scheib S, Patzkowsky K (2016) Advantages \nof robotics in benign gynecologic surgery. Curr Opin Obstet \nGynecol 28(4):304–310\n12. Sinha R, Rallabandi H, Bana R, Bag M, Raina R, H K D DS, \nReddy P (2024) Ovarian loss in laparoscopic and robotic cys -\ntectomy compared using artificial intelligence pathology. JSLS \n28(1):e2024.00001\n13. Evans AJ, Brown RW, Bui MM, Chlipala EA, Lacchetti C, Mil -\nner DA et al (2022) Validating Whole Slide Imaging Systems \nfor Diagnostic Purposes in Pathology. Arch Pathol Lab Med \n146(4):440–450\n14. Broer SL, Broekmans FJ, Laven JS, Fauser BC (2014) Anti-Mül-\nlerian hormone: ovarian reserve testing and its potential clinical \nimplications. Hum Reprod Update 20(5):688–701\n15. Urman B, Alper E, Yakin K, Oktem O, Aksoy S, Alatas C et al \n(2013) Removal of unilateral endometriomas is associated with \nimmediate and sustained reduction in ovarian reserve. Reprod \nBiomed Online 27(2):212–216\n16. Jiang D, Nie X (2020) Effect of endometrioma and its surgical \nexcision on fertility (Review). Exp Ther Med 20(5):114\n17. Coelho Neto MA, Ludwin A, Borrell A, Benacerraf B, Dewailly \nD, da Silva Costa F et al (2018) Counting ovarian antral folli -\ncles by ultrasound: a practical guide. Ultrasound Obstet Gynecol \n51(1):10–20\n18. Somigliana E, Ragni G, Benedetti F, Borroni R, Vegetti W, Cro-\nsignani PG (2003) Does laparoscopic excision of endometriotic \novarian cysts significantly affect ovarian reserve? Insights from \nIVF cycles. Hum Reprod 18(11):2450–2453\n19. Ragni G, Somigliana E, Benedetti F, Paffoni A, Vegetti W, Restelli \nL et al (2005) Damage to ovarian reserve associated with laparo-\nscopic excision of endometriomas: a quantitative rather than a \nqualitative injury. Am J Obstet Gynecol 193(6):1908–1914\n20. Murdock C, Sanchez-Ramos L, McKinney JA, Carrubba AR, \nLewis G (2025) The Impact of Laparoscopic Cystectomy for \nOvarian Endometrioma on Anti-Müllerian Hormone Levels: A \nSystematic Review and Meta-Analysis. Gynecol Obstet Invest \n90(6):657–671\n21. Yu HT, Huang HY , Tseng HJ, Wang CJ, Lee CL, Soong YK \n(2017) Bilaterality of ovarian endometriomas does not affect the \noutcome of in vitro fertilization/intracytoplasmic sperm injec -\ntion in infertile women after laparoscopic cystectomy. Biomed J \n40(5):295–299\n22. Younis JS, Shapso N, Ben-Sira Y , Nelson SM, Izhaki I (2022) \nEndometrioma surgery-a systematic review and meta-analysis \nof the effect on antral follicle count and anti-Müllerian hormone. \nAm J Obstet Gynecol 226(1):33–51e7\n23. Asgari Z, Rouholamin S, Hosseini R, Sepidarkish M, Hafizi L, \nJavaheri A (2016) Comparing ovarian reserve after laparoscopic \nexcision of endometriotic cysts and hemostasis achieved either by \nbipolar coagulation or suturing: a randomized clinical trial. Arch \nGynecol Obstet 293(5):1015–1022\n24. Lee CL (2024) Trends in the Surgical Treatment of Benign Uter-\nine Tumors: Conservative vs. Radical Approaches. Gynecol \nMinim Invasive Ther 13(3):139–140\n25. McDonnell R, Gollow J, Nathan E, Doherty D, Majumder K, \nWilkinson E et al (2025) Endometriosis Quality of Life Cohort \nStudy: Long-term Impact of Radical Laparoscopic Excision of \nEndometriosis. Gynecol Minim Invasive Ther 14(1):57–65\n26. Sztyler-Krąkowska M, Wąsowicz A (2026) Clinical Implications \nof Extracellular Vesicles in Endometriosis: A Systematic Review. \nGynecol Minim Invasive Ther 15(1):65–74\nPublisher’s note Springer Nature remains neutral with regard to juris-\ndictional claims in published maps and institutional affiliations.","source_license":"CC0","license_restricted":false}