Vaginal natural orifice transluminal endoscopic surgery for malignant ovarian tumors: a single-institution study.

This was a retrospective single-institution cohort of consecutive patients who underwent vNOTES for malignant ovarian conditions from May 2021 to September 2024. All patients included in this study had histology-proven malignancy arising from or involving the ovary. Centralized Institutional Review Board ethics approval was obtained for the use of anonymized patient data. Prior to the surgery, all patients were investigated pre-operatively by both surgical and anesthetic teams. Imaging investigations could include transvaginal and/or transabdominal pelvic ultrasound scans, computed tomography (CT), and magnetic resonance imaging (MRI) of the pelvis. Bloodwork included ovarian tumor markers, such as CA-125, CA-19-9, beta-HCG, alfa-fetoprotein, and carcinoembryonic antigen. Patients were admitted on the same day of the surgery. Procedures were carried out under general anesthesia. All patients did not receive bowel preparation prior to surgery. Prophylactic intravenous (IV) cefazolin and metronidazole (or clindamycin in the case of documented penicillin or cephalosporin allergy) were given prior to induction and within one hour of incision. The patient was placed in a lithotomy position, cleaned, draped and catheterized. The cervical lip was grasped with vulsellum forceps, and a circumferential incision around the cervix was made. In cases of staging, restaging or interval debulking surgery, sharp dissection was used to perform an anterior and a posterior colpotomy to gain entry into the intraperitoneal pelvic cavity. In cases of fertility-sparing surgery, only a posterior colpotomy was performed for intraperitoneal access. Thereafter, the uterosacral and cardinal ligaments are identified and sealed with an advanced biopolar device such as Enseal (Ethicon, Rarican, NJ, USA). An extra-extra-small (XXS) Alexis (Applied Medical, Rancho Santa Margarita, CA, USA) wound retractor was inserted into the Pouch of Douglas (POD) and vesicovaginal space as a platform for port placement. At our institution, we adopted the use of a home-made glove port consisting of a sterile size 7.5 left-hand glove attached to the Alexis retractor as depicted in the figure below (Fig.  1 ). Three 5 mm ports and one 12 mm plastic ports were fitted into the little finger, middle finger, index finger and thumb sleeves respectively and secured with silk sutures to prevent air leak. Pneumoperitoneum was established with 12 mmHg of carbon dioxide. A 45 mm or 50 mm 30-degree rigid long laparoscope was introduced. A systematic peritoneal survey was performed to assess the peritoneum, omentum, bowel, diaphragmatic surface, liver surface, and stomach for evidence of intracoelomic metastasis. Peritoneal washings were taken for cytology. Where performed, transperitoneal or retroperitoneal pelvic lymph node dissection (PLND) was performed to the level of the common iliac artery. Frozen section was performed in cases where ovarian masses were suspicious for cancer. If positive for malignancy, THBSO with PLND and omentectomy was performed; borderline tumors on the other hand received THBSO and omentectomy. Fig. 1 vNOTES glove method setup vNOTES glove method setup In our study, we did not encounter any cases of tumor spillage during specimen retrieval. All specimens were placed into a surgical bag prior to vaginal retrieval to avoid intraperitoneal spillage and potential upstaging of cancer. After ensuring hemostasis, the Alexis retractor was removed together with the glove port. The vaginal vault was closed with Vicryl 2-0 sutures. Informed consent was obtained from all patients for the vNOTES procedure, with detailed explanation of the potential anesthetic and surgical risks of the procedure. All patients in our study cohort underwent preoperative imaging. In cases where there was a suspicion of ovarian malignancy, patients were counseled and provided informed consent for the possibility of an intraoperative frozen section. As part of this process, they were informed about the limitations and accuracy of the frozen section technique, and the potential need to proceed with a comprehensive staging procedure, such as omentectomy and/or pelvic lymph node dissection, based on the frozen section results. All patients were managed post-operatively according to recommendations from the Enhanced Recovery After Surgery (ERAS) protocol consisting of regular analgesia, early initiation of oral feeding, early ambulation, and early removal of the indwelling catheter post-operatively. Background characteristics, intraoperative outcomes, postoperative complications, and histopathological results were collected from electronic medical records and tabulated. Patients were classified to one of four categories by operative indication: Primary staging surgery Restaging surgery Fertility-sparing surgery Interval debulking surgery (after chemotherapy) Primary staging surgery Restaging surgery Fertility-sparing surgery Interval debulking surgery (after chemotherapy) As the primary staging surgery and fertility-sparing surgery group had the largest number of patients, they were summarized by tabulation. Only 1 patient underwent restaging surgery and 2 patients underwent interval debulking surgery and they were descriptively analyzed. Shapiro–Wilk test was used to assess for normality. Continuous variables were expressed as median and interquartile range, while categorical variables were expressed as absolute numbers and percentages. Surgical complications encountered were classified according to the Clavien–Dindo Classification. Kaplan–Meier curve for tumor recurrence was plotted using the dates of latest follow-up. Descriptive analyses of patients who had experienced tumor recurrence were also performed. All statistical analyses were conducted in R-4.3.0.

A total of 19 patients with malignant tumors arising from or involving the ovary were included for analysis. 4 patients had fertility-sparing surgery, 12 patients had primary staging surgery, 1 patient underwent restaging surgery and 2 patients underwent interval debulking surgery (Table  1 ). Table 1 Baseline demographics of patients included in the study Fertility-sparing ( N  = 4) Primary staging ( N  = 12) Age 37 [33, 38] 62 [54, 67.5] BMI 21 [20.6, 24.2] 25.9 [22.7, 30.0] Hypertension 1 (25.0) 6 (50.0) Hyperlipidemia 0 7 (58.3) Diabetes mellitus 0 3 (25.0) History of breast cancer 0 3 (25.0) Parity (%)  0 2 (50.0) 3 (25.0)  1 1 (25.0) 0  ≥2 1 (25.0) 9 (75.0) Previous abdominal surgery 2 (50.0) 5 (41.7) Tumor size, cm 14 [7.6, 21] 3.3 [2.1, 5.2] Expressed as median [interquartile range] or n (%) BMI body mass index Baseline demographics of patients included in the study Expressed as median [interquartile range] or n (%) BMI body mass index The median age was 62 in the staging surgery group and 37 in the fertility-sparing surgery group. More patients in the primary staging surgery group had comorbidities of hypertension, hyperlipidemia and diabetes as compared to only 1 patient with hypertension in the fertility-sparing surgery group. Tumor size was noticeably larger in the fertility-sparing group as many patients in this group initially presented as seemingly benign large ovarian cysts. In the primary staging surgery group, hysterectomy was performed in all patients, omentectomy in a third of the patients, and PLND in a quarter of the patients (Table  2 ). For the two-thirds of patients who did not undergo omentectomy, intraoperative frozen section did not show malignancy but this was present on final histology. Only one patient in our series experienced an intraoperative complication of high EBL (1000 mL). This was a 49-year-old patient with synchronous ovarian and endometrial endometrioid carcinoma; she was a known hemophilia carrier who also had a previous history of postpartum hemorrhage. Surgical-site bleeding was the cause of high EBL and was treated with transfusion of 1 pint of packed red blood cells and 2 units of fresh frozen plasma. Pain scores were mostly zero at 12 and 24 h post-operatively, and most patients were discharged on postoperative day 1 or 2. There were no readmissions for postoperative complications or disease recurrence within 30 days. Median follow-up time was 26.4 months in the whole cohort (interquartile range, 6.3 to 30 months). Table 2 Operative and postoperative outcomes Fertility-sparing ( N  = 4) Primary staging ( N  = 12) Ovarian operation Bilateral oophorectomy 0 12 (100) Unilateral oophorectomy 2 (50.0) 0 Unilateral cystectomy 2 (50.0) 0 Accompanying procedure  Hysterectomy 0 12 (100)  Omentectomy 1 (25.0) 4 (33.3)  PLND 0 3 (25.0) Operative time, min 105 [91, 108] 135 [114, 221] EBL, mL 75 [50, 125] 100 [100, 200] Intraoperative complications 0 1 (8.3) VAS pain score  4 h post-operatively 2 [1.5, 2.25] 0 [0, 0.75]  8 h post-operatively 0 [0, 0] 0 [0, 0.5]  12 h post-operatively 0 [0, 1] 0 [0, 0]  24 h post-operatively 0 [0, 0.25] 0 [0, 0] Length of stay, days 2 [2] 2.5 [2, 3] Tumor type  Ovarian only 4 (100) 7 (58.3)  Ovarian + endometrial 0 3 (25.0)  Ovarian + fallopian tube 0 2 (16.7) Peritoneal washings positive 0 2 (16.7) Omental biopsy positive 0 1 (8.3) Postoperative chemotherapy  Indicated and received 1 (33.3) 4 (33.3)  Indicated but declined by patient 0 4 (33.3)  Not indicated 2 (66.7) 4 (33.3) Follow-up time, months 11.7 [7.1, 17.7] 27.6 [7.8, 31.4] Expressed as median [interquartile range] or n (%) EBL estimated blood loss, PLND pelvic lymph node dissection, VAS Visual Analog Scale Operative and postoperative outcomes Expressed as median [interquartile range] or n (%) EBL estimated blood loss, PLND pelvic lymph node dissection, VAS Visual Analog Scale Fertility-sparing surgery was performed in 4 patients. The first was a 37-year-old who underwent vNOTES left ovarian cystectomy for a 20 cm ovarian cyst with normal tumor markers and no suspicious features on preoperative scans. Histology reveled granulosa cell tumor, and she later underwent restaging fertility-sparing vNOTES salpingo-oophorectomy and omentectomy. Final histology showed no residual tumor and the patient was counseled for chemotherapy but declined, and is disease-free at 20.5 months of follow-up. The second was a 37-year-old who initially underwent conventional laparoscopic ovarian cystectomy for a 20 cm ovarian cyst. Histology revealed ovarian dysgerminoma and she underwent restaging vNOTES left salpingo-oophorectomy and infracolic omentectomy with no evidence of residual tumor on histology. She was counseled for chemotherapy but declined, and is disease-free at 15 months of follow-up. The third was a 40-year-old who underwent vNOTES left salpingo-oophorectomy and adhesiolysis for an ovarian mass showing malignant struma ovarii on histology. She was disease-free at 8.4 months at follow-up. The fourth was a 21-year-old with incidental finding of a seemingly benign ovarian cyst on ultrasound and normal ovarian tumor markers; she underwent vNOTES left ovarian cystectomy. The final histology reveled Sertoli–Leydig cell tumor, and she underwent restaging vNOTES left salpingo-oophorectomy. After the restaging surgery, histology showed no residual disease and she completed 3 cycles of adjuvant BEP chemotherapy. The one patient who underwent a restaging surgery was a 47-year-old who initially underwent emergency laparotomy with right salpingo-oophorectomy for abdominal pain, revealing hemoperitoneum and a ruptured tumor. Histological analysis of the right ovary revealed a granulosa cell tumor measuring around 20 cm in size. She then underwent restaging surgery in the form of vNOTES total hysterectomy, left salpingo-oophorectomy and omentectomy, with final histology confirming stage 3A2 adult granulosa cell tumor of the ovary and involving the omentum. She underwent postoperative chemotherapy but developed disease recurrence in the peritoneum 19.6 months after surgery. This was addressed with exploratory laparotomy, palliative debulking surgery, low anterior resection and defunctioning double-barrel ileostomy, and the patient was last noted to be on best supportive care at 35.8 months post-vNOTES surgery. Two patients underwent interval debulking surgery. The first was a 58-year-old BRCA carrier with stage 3C ovarian cancer with extensive carcinomatosis peritonei on the initial scan. After 3 cycles of neoadjuvant chemotherapy, there were no large tumors seen on CT scan and the patient underwent interval vNOTES THBSO and omentectomy. Intraoperatively, the peritoneal survey was normal. Histology revealed microscopic foci of synchronous high-grade serous carcinoma of the ovarian surface, fallopian tube and omentum, associated with serous tubal intraepithelial carcinoma of the fallopian tube. She developed metastatic recurrence 18 months post-operatively, presenting as a malignant right pleural effusion and raised CA-125 although without peritoneal disease. Cytology of the pleural fluid confirmed metastatic fallopian tube tumor, and the patient is currently on palliative chemotherapy. The second patient was a 75-year-old with multiple medical comorbidities (including tuberculosis and end-stage renal failure) who underwent interval vNOTES THBSO, right pelvic peritonectomy, and omentectomy. She was initially treated with 3 cycles of neoadjuvant chemotherapy before surgery. Peritoneal surgery was clear intraoperatively. Final histology showed high-grade serous carcinoma of right tubal primary, with microscopic involvement of the ovary and omentum. The patient was discharged well without renal complications and is currently undergoing adjuvant chemotherapy 1.25 months post-operatively. Recurrence occurred in 4 patients during follow-up and was picked up by rising ovarian tumor marker levels and/or repeat scans. No deaths were reported across follow-up. Kaplan–Meier curve for recurrence is shown in Fig.  2 . Details of patients with a recurrent tumor are as follows: This patient underwent primary staging vNOTES THBSO for Stage 1C3 synchronous ovarian and endometrial cancer tumor. The tumor board recommended adjuvant chemotherapy due to positive peritoneal cytology and ovarian surface involvement but the patient declined. Subsequently the patient had a left adnexal recurrence at 24.7 months. This was then treated with laparotomy, resection of pelvic tumor and left internal iliac artery, debulking and adhesiolysis. Histology from the second operation revealed adenocarcinoma in keeping with recurrent endometrioid carcinoma of the ovary, and she received 6 cycles of paclitaxel and carboplatin chemotherapy. At 40.6 months of post-vNOTES surgery, she was alive and recurrence-free. This was a BRCA2 patient who underwent prophylactic vNOTES THBSO and thereafter underwent completion laparoscopic PLND and omentectomy which showed no residual disease. She then had adjuvant chemotherapy. She had a recurrence in the pelvis at 20.4 months which was confirmed high-grade serous ovarian carcinoma. She was treated with 3 further cycles of chemotherapy with partial response and is currently on maintenance olaparib. The 47-year-old patient who underwent restaging surgery as mentioned earlier had a recurrence in the peritoneum 19.6 months post-operatively. The 58-year-old patient who underwent interval debulking surgery as mentioned earlier had a metastatic recurrence presenting as pleural effusion at 19.2 months post-operatively. Fig. 2 Kaplan–Meier curve of recurrence This patient underwent primary staging vNOTES THBSO for Stage 1C3 synchronous ovarian and endometrial cancer tumor. The tumor board recommended adjuvant chemotherapy due to positive peritoneal cytology and ovarian surface involvement but the patient declined. Subsequently the patient had a left adnexal recurrence at 24.7 months. This was then treated with laparotomy, resection of pelvic tumor and left internal iliac artery, debulking and adhesiolysis. Histology from the second operation revealed adenocarcinoma in keeping with recurrent endometrioid carcinoma of the ovary, and she received 6 cycles of paclitaxel and carboplatin chemotherapy. At 40.6 months of post-vNOTES surgery, she was alive and recurrence-free. This was a BRCA2 patient who underwent prophylactic vNOTES THBSO and thereafter underwent completion laparoscopic PLND and omentectomy which showed no residual disease. She then had adjuvant chemotherapy. She had a recurrence in the pelvis at 20.4 months which was confirmed high-grade serous ovarian carcinoma. She was treated with 3 further cycles of chemotherapy with partial response and is currently on maintenance olaparib. The 47-year-old patient who underwent restaging surgery as mentioned earlier had a recurrence in the peritoneum 19.6 months post-operatively. The 58-year-old patient who underwent interval debulking surgery as mentioned earlier had a metastatic recurrence presenting as pleural effusion at 19.2 months post-operatively. Kaplan–Meier curve of recurrence

Our study represents the largest cohort of patients undergoing vNOTES for ovarian malignancy in current literature. Low rates of intraoperative and postoperative complications, short length of stay, and favorable short- to medium-term oncological outcomes were seen with this technique.

Ovarian cancer surgery was historically done via an open approach, but minimally invasive surgery (MIS), such as laparoscopy and robotic surgery, has also proven to be safe and effective [ 16 , 17 ]. MIS has shown feasibility across various stages of ovarian cancer, ranging from surgical staging in small, early-stage tumors to cytoreductive surgery in advanced cases. A case–control study demonstrated that survival outcomes are not significantly affected by the type of surgical approach, be it laparoscopy, robotic, or laparotomy [ 18 ]. A meta-analysis also showed similar recurrence rates between laparoscopy and laparotomy for apparent early-stage ovarian tumors [ 19 ]. Nonetheless, appropriate patient selection remains a prerequisite for good oncological outcome without necessitating conversion to open surgery. Within the small operating field and smaller incisions utilized, there are concerns related to tumor rupture causing intraperitoneal dissemination, difficulty in removing the adnexal mass intact without spillage, adequate retroperitoneal staging, and fertility-sparing surgery [ 20 ]. Similar concerns exist with vNOTES surgery, which can be thought of as laparoendoscopic single-site (LESS) surgery with colpotomy as the single-site incision. Although vNOTES and LESS offer advantages of only requiring a single incision and better cosmesis over standard laparoscopic surgery, the cramped entry space for instruments, in almost the same plane as the endoscopic camera, may hinder intra-peritoneal mobilization, obscure camera vision, and contribute to instrument clashing even in the most experienced hands. These concerns are particularly important for oncological debulking or restaging surgeries wherein complete, maximum cytoreduction is desired. Hence, although vNOTES has received exponential adoption as an effective procedure for benign gynecological conditions, such as leiomyomas, endometriosis and ovarian cysts [ 21 – 24 ], it has yet to receive widespread adoption for malignant conditions. Nonetheless, preliminary studies of vNOTES for malignant gynecological conditions have shown much promise. In the setting of endometrial cancer, favorable short-term outcomes were noted, with decreased length of stay and superior cosmesis to standard laparoscopy while also having a low rate of postoperative complications. Sentinel lymph node mapping is also achievable at a similar success rate to standard laparoscopy [ 25 ]. In the setting of ovarian cancer, evidence is sparse but still appears favorable. Lowenstein et al. performed ovarian cancer surgery with omentectomy in 5 patients via vNOTES using a LESS trocar, with comparable omentectomy time to conventional laparoscopy and laparotomy [ 26 ]. Hurni et al. described the largest cohort in literature so far of 14 patients with ovarian tumors, of which 7 had borderline or malignant histology and 7 had benign histology [ 15 ]. These patients mostly underwent THBSO with infracolic omentectomy and peritoneal washings as part of primary staging surgery, with a smaller proportion undergoing unilateral salpingo-oophorectomy instead of THBSO. There were no intraoperative complications and only 1 case of postoperative fever associated with vaginal cuff infection. Besides the use of glove ports instead of LESS ports for vaginal access, the surgical technique in our study is similar to that reported by Hurni et al. Favorable postoperative outcomes including length of stay and pain scores were observed in both studies. Patient selection is also an important consideration for vNOTES despite its utility in various scenarios, such as obesity and previous abdominal surgery. The main reasons for avoiding vNOTES in our institution are a history of severe endometriosis or previous anterior resection, due to the increased likelihood of dense adhesions at the Pouch of Douglas. Additionally, patients with tumors deemed unsuitable for vaginal retrieval or those considered to carry a high risk of intraoperative tumor spillage were not selected for the vNOTES approach. Our cohort of patients notably also includes patients who underwent fertility-sparing vNOTES. As described earlier, there were two cases of vNOTES cystectomy for what initially seemed to be benign ovarian cysts on imaging with normal ovarian tumor markers. After malignant histology was noted, vNOTES restaging surgery with salpingo-oophorectomy was also feasible and revealed no residual disease in both patients. To prevent tumor spillage, in both cases, the ovary with its large cyst was delivered via the vagina, and cystectomy was done in an extracorporeal fashion. These cases demonstrate the versatility of vNOTES technique even for technically challenging large cysts. This study also adds to literature with reporting of short- to medium-term oncological outcomes. No deaths were reported within this cohort of 19 patients, and only 4 patients experienced a recurrence. In the case of the adnexal recurrence after 24.7 months, intraperitoneal cytology was already positive and the ovarian surface was already involved; with the patient also declining chemotherapy, recurrence was an expected event. Similarly, the patient presenting with ruptured tumor and undergoing vNOTES restaging surgery was at high risk of recurrence due to having pre-existing tumor seeding from the rupture. For the other two cases (pelvic and lung recurrence), there was no evidence of residual disease after surgery, and the recurrences appear much more likely due to aggressive tumor biology rather than incomplete resection. Although longer-term survival and recurrence outcomes in larger cohorts are required to perform robust statistical analyses comparing vNOTES against other forms of MIS and open surgery, these early results are encouraging. Further directions include investigation of the robot-assisted vNOTES approach which has been performed in the United States and China with encouraging results [ 27 , 28 ]. Several limitations are acknowledged. Despite being the largest cohort of vNOTES for ovarian malignancy to our knowledge, this study still included a relatively small number of patients and future studies are needed to cement vNOTES as a keystone technique for ovarian malignancy. There was an insufficient number of patients to perform univariate or multivariate analyses to elucidate the optimal patient selection characteristics or factors predicting recurrence. All cases in this study were performed by a senior surgeon with experience of over 400 vNOTES cases to date, as well as extensive experience in minimally invasive surgery, such as conventional laparoscopy, LESS, and robotic surgery. The learning curve for this technique was not analyzed; recent study showed that the learning curve ranges from 7 to 16 cases [ 29 ]. As mentioned earlier, longer-term follow-up will be beneficial to establish oncological equivalence to other surgical approaches.

Vaginal natural orifice transluminal endoscopic surgery (vNOTES) is an established access technique for gynecological surgeries [ 1 ]. vNOTES hysterectomy has shown similar operative outcomes to laparoscopic hysterectomy for benign indications with the notable advantages of reduced length of stay (LOS), lower pain scores, and lower estimated blood loss [ 2 – 5 ]. Other applications of vNOTES include salpingectomy, myomectomy, ovarian cystectomy, urogynecological surgery, colorectal surgery, cholecystectomy, and donor nephrectomy [ 1 , 6 – 9 ]. This technique has also been utilized effectively in the setting of obesity, including the super-morbidly obese [ 10 ], and can be safely performed even in patients with large uteri [ 11 ]. vNOTES is increasingly being used to treat malignant conditions such as endometrial cancer; staging surgeries with pelvic lymph node dissection (PLND) and omentectomy can also be safely performed [ 12 , 13 ]. However, its application to ovarian cancer is not well-characterized at present, with most studies focusing on cystectomy for benign ovarian cysts [ 14 ]. The only study with a sizable cohort of malignant ovarian tumors is a recent study by Hurni et al. which focused on omentectomy for suspicious adnexal masses and high-risk endometrial cancers, proving the feasibility of vNOTES oncological staging [ 15 ]. In lieu of the paucity of literature describing vNOTES for malignant ovarian tumors, we aimed to describe the perioperative and medium-term oncological outcomes of this procedure in our institution.