Pneumovaginoscopy-assisted radical hysterectomy for early-stage cervical cancer: a novel bidirectional approach for tumor spillage prevention and R0 resection.

In the recent National Comprehensive Cancer Network Guidelines [ 1 ], the indication for radical hysterectomy for cervical carcinoma is stage IA1 with lymphovascular space invasion (LVSI), stage IB1–3, and stage IIA2. Historically, there are subtypes of radical hysterectomy [ 2 3 ]. They vary in terms of the technical challenge, the extent and radicality of tumor resection, and invasiveness, especially to nerves around the bladder. The modern classification of radical hysterectomy is based on the extent of paracervical resection [ 4 ]. According to this classification, radical hysterectomy type B2 (transection of the paracervix at the ureter with the removal of lateral paracervical nodes) is usually appropriate for cancer stages IB1 and IB2, without bladder nerve damage. However, adjustment for the vaginal cuff length is limited technically. Furthermore, tumor spillage is the current issue delaying pursuits to implement minimally invasive radical hysterectomy for early-stage cervical carcinoma [ 5 6 ]. Vaginal cuff creation and vaginal irrigation have been suggested to prevent tumor spillage [ 6 7 ]. On the other hand, transvaginal natural orifice transluminal endoscopic surgery (vNOTES) has recently been performed as a gynecological vaginal endoscopy [ 8 ]. Moreover, we developed a pneumovaginoscopic surgical technique using the same device and similar technique to perform precise endoscopic vaginal surgery [ 9 10 ]. Regarding its application for early-stage cervical carcinoma, we hypothesized that this technique could enable the accurate diagnosis and resection of vaginal invasion, secure vaginal cuff creation, and allow irrigation to prevent recurrence caused by tumor spillage. In addition, pneumovaginoscopy (PV) should enable closer inspection of the anatomy, especially the endopelvic fascia around the cervix, and more targeted resection of the vaginal aspect of the paracervix (paracolpium) to simplify nerve-sparing radical hysterectomy. Therefore, the present study aimed to evaluate the feasibility and potential of this surgery.

The clinical characteristics of the study patients and treatment categories are shown in Table 1 . Thirty-eight (64.4%) patients had stage IB1 cancer. Seven patients (11.9%) had vaginal invasion (4 cases of stage IIA1 and 3 cases of stage IIA2). Ten (16.9%) patients were obese with a body mass index of ≥25 kg/m 2 . Twenty-one patients (35.6%) underwent laparoscopic surgeries. Twenty-three (39.0%) received adjuvant chemotherapy for lymph node metastasis (n=8), vessel infiltration (n=10), or lymph vessel infiltration (n=7), as shown in Table 2 . One patient received adjuvant CCRT because the histology revealed gastric-type adenocarcinoma. Data shown are number (%) not otherwise specified. BMI, body mass index; CCRT, concurrent chemoradiotherapy; SCC, squamous cell carcinoma; SD; standard deviation. * Gastric type adenocarcinoma. Values are shown as median (range) or number (%) as appropriate. Operative outcomes are summarized in Table 2 . The median (range) operative time was 346 (191–608) minutes. The estimated median (range) blood loss was 281 (0–2,632) mL. There were three (5.1%) intraoperative complications: CO 2 gas embolism (n=1) and sigmoid colon muscle layer injury (n=1) in stage IB1 patients, and ureteral injury (n=1) in stage IIA2 patients. CO 2 gas embolism occurred while ensuring hemostasis of a vaginal vein injury during PV; the patient fully recovered during the procedure without postoperative repercussions (discussed below). A sigmoid colon injury occurred during the laparoscopy of an obese patient and was repaired during the operation. A ureteral injury occurred during surgery on a patient staged as IIA2 with vaginal invasion; the injury was repaired during the procedure without postoperative complications. There were 8 (16.3%) postoperative complications. Six occurred in patients with stage IB1 cancer: lymphedema with cellulitis (n=3), vaginal cuff dehiscence (n=1), sub-ileus (n=1), and symptomatic lymphocyst (n=1); two occurred in patients with stage IIA1 cancer: lymphedema with cellulitis (n=1) and ureterovaginal fistula (n=1). The mean (range) period to recover residual urine volume <50 mL was 3 (3–336) days. One patient experienced a long delay (336 days) in recovering voiding function (discussed below). This procedure achieved microscopic R0 at both vaginal and parametrial margins in all cases. Survival outcomes according to the Kaplan–Meier analysis of DFS and OS for all patients are shown in Fig. 4 and Table S1 . The median follow-up was 44.5 (2–122) months, and no recurrence was observed. DFS, disease-free survival; OS, overall survival.

In this report, we found that this surgical technique with PV was feasible for early-stage cervical carcinoma. It showed potential for accurate diagnosis and resection of the vaginal invasion and enabled secure vaginal cuff creation and irrigation to avoid tumor spillage. Moreover, the pneumovaginoscopic inspection and resection of the vaginal aspect of the paracervix (paracolpium) provided better visualization of the surgical anatomy for the fascia-oriented and nerve-sparing radical hysterectomy. PVRH was feasible for early-stage cervical carcinoma up to stage IIA. All procedures were performed within appropriate metrics for operation time and blood loss. This surgery was recently performed on patients with stage 2B and 3B cervical carcinoma after neoadjuvant chemotherapy. Including these, more than 125 cases of PVRH have been performed at our institution. There were no conversions to conventional radical hysterectomy. The rates and types of intraoperative and postoperative complications differed little from those of conventional procedures, and all patients recovered completely either peri- or postoperatively. Laparoscopic surgery was recommended to the patients as minimally invasive surgery, whereas open surgery was primarily conducted for an enlarged uterus (including fibroids), contraindications to laparoscopy (including glaucoma), and the preference of the patient. Moreover, open and laparoscopic surgery differ in total operation time, total blood loss, and duration of hospital stay, as previously described. However, this did not substantially affect the PV procedure. This surgery showed the potential to enable accurate evaluation and resection of vaginal invasion. In all cases, the cervical and vaginal lesions of the cervical carcinoma were observed clearly with the pneumovaginoscope, and the vaginal margins were pathologically negative, even in cases with vaginal invasion up to stage IIA. Positive surgical margin has been considered one of the most critical prognostic risk factors [ 1 ]. In addition, the Schiller test has been used and continues to be one of the most useful means of improving the detection of cervical malignancies [ 12 ]. We have also reported successful pneumovaginoscopic resection of vaginal malignancies [ 9 10 ]. Endoscopically enhanced direct view might be one of the best methods to observe vaginal lesions [ 13 ]. Excision of vaginal areas under direct observation might contribute to R0 resection and nonrecurrence. A second benefit of this surgical technique is the potential for easy vaginal cuff creation and irrigation to avoid tumor spillage. In our cohort, vaginal cuffs were created, and irrigation was performed successfully in all cases, without postoperative peritoneal dissemination over follow-up, as reported by Ramirez et al. [ 5 ]. Kanao et al. [ 6 ] recommended vaginal cuff creation to prevent tumor spillage and reported excellent outcomes. A similar concept to the PVRH approach is the trans-anal total mesorectal excision (TaTME) for rectal cancer. However, this approach has a risk of tumor spillage, and its use is being phased out in Europe for this reason. Loose closure has been suggested as the cause [ 14 15 ]. Therefore, we used barbed sutures during vaginal cuff creation to ensure the sutures did not loosen. The water-tightness of barbed suture closure has been demonstrated compared with conventional sutures [ 16 ]. We also irrigated the vaginal canal thoroughly with normal saline and disinfected it with isodine after vaginal cuff creation. This procedure might also have contributed to preventing tumor spillage and, thus, maintained the absent/low recurrence rate. Washout is an established technique to minimize contamination [ 17 ]. In addition, several studies have supported the use of irrigation with normal saline and isodine disinfection to reduce the viability of disseminating tumor cells [ 18 19 ]. Positive CO 2 gas pressure has been hypothesized to enhance the dissemination or metastasis of cancerous cells [ 20 ], including during laparoscopic radical hysterectomy [ 21 ]. However, meta-analyses have also contradicted this theory [ 22 ]. Regardless, in our cohort, it appeared that vaginal cuff creation isolated the cancer cells physically and practically, vaginal irrigation washed out the isolated cancer cells, and isodine application reduced the viability of any residual cancer cells. These procedures were likely contributors to the lack of postoperative recurrence. The third obvious benefit of the pneumovaginoscopic technique was the ability to evaluate the vaginal paracolpium prior to resection. This enhanced visualization of the surgical anatomy and simplified nerve-sparing radical hysterectomy. As evident in Fig. 2 , PV enabled visualization and resection of the paracolpium (vaginal vessels and surrounding connective tissues) to the cervix, just below the level of the vesicovaginal septal attachment, from the vaginal side and without the need for direct identification and isolation of the bladder nerves. The urinary recovery time was comparable to that of other nerve-sparing radical hysterectomies [ 6 23 ]. The recent anatomical clarification of these nerves and the surrounding blood vessels have established the role of nerve-sparing techniques when performing radical hysterectomy [ 24 25 ]. However, isolation and resection of these vessels, ligaments, and nerves is sometimes difficult and can cause troublesome bleeding and thermal damage, mainly because of the anatomical complexity and the effects of cancer invasion. PV enables direct vaginal dissection of the paracolpium and can minimize the need for more invasive maneuvers. PV also enabled the isolation of the uterine fascia (vesicovaginal and rectovaginal fascia) and lower uterine ligaments (including the vesicovaginal, cardinal, and sacrouterine ligaments) from the dorsal side as a single structure (PCEPF) surrounding the cervix. It also enabled the circular incision of this PCEPF while excising the entire cervix and corpus with the blood vessels and surrounding fascia within a visceral fascial envelope that was still intact ( Fig. 3 ). PCEPF could be the membranous landmark that enables the separation of early-stage cervical cancer from pelvic nerves intact, similar to TaTME [ 26 ] and radical prostatectomy [ 27 ]. The isolation of PCEPF might be important for other gynecologic surgeries. The anterior surface of the PCEPF is also isolated during conventional total laparoscopic hysterectomy. This occurs after resection of the surrounding peritoneum and uterine vessels when the vaginal fornices delineator is used to push up the vaginal fornix ( Fig. 2K ). This isolation confirms the ureteral separation from the cervix and the extrafascial hysterectomy, which is critical, especially for endometrial cancer. We also found that the rectum and rectovaginal septum were easily separated using pneumovaginal gas pressure, even in the case of severe endometriosis. Moreover, the dorsal isolation of the sacrouterine ligaments might be helpful, especially for treating deep infiltrating endometriosis. We encountered one case of CO 2 gas embolism during surgery (case #9). Gas embolism is a rare but occasionally serious complication during laparoscopy [ 28 ]. The risk of gas embolism in TaTME, which requires trans-anal insufflation of CO 2 and displays certain procedural similarities to our pneumovaginoscopic surgery, has also been reported recently [ 29 ]. Our embolism occurred during bleeding and hemostasis of the paracolpial venous plexus in an early case from this study. Here, pneumatic pressure was elevated temporarily from 10 to 15 mmHg to tamp venous bleeding; this is the suspected cause of this complication. The embolism was recovered during the procedure, and the operation was completed. After this episode, the anatomy of the vaginal vessels was gradually clarified by PV, as shown in Fig. 1 , which prevented vessel bleeding. The pressure level was maintained at <14 mmHg. Since the refinement of this procedure, no such complications have occurred. Two ureter complications occurred, 1 intraoperatively (case #20) and 1 postoperatively (case #26). Both occurred in patients with stage IIA advanced cervical carcinoma. They recovered completely and avoided postoperative radiotherapy, which might have led to postoperative ureteral stenosis in such a situation. However, ureteral complications from this type of surgery for vaginal invasion are possible. One patient experienced a long delay before urination recovered (case #12). Additionally, this patient was an early case of this study, and the lateral fascia was accidentally dissected during the paracolpium resection. This complication suggests that the lateral fascia was just adjacent to the hypogastric nerve and was equivalent to the pre-hypogastric nerve fascia [ 25 26 ]. All of the above complications occurred in the first half of this study. After clarifying the local anatomy and establishing the technique, no notable complications occurred in the second half of the study and in the recent cases of PVRH for patients with stages 2B and 3B cervical carcinoma after neoadjuvant chemotherapy, of which there are now more than 125. This study comprised a relatively small number of cases without a comparative control group. Moreover, the study was performed by a single surgeon at two institutions and had a limited follow-up time. A confirmatory study, including long-term prognostic studies from multiple facilities, is expected. In conclusion, PV was feasible for early-stage cervical carcinoma. It can be considered a potential technique for accurate and less-invasive radical hysterectomy. PV during radical hysterectomy revealed a new surgical anatomy, which shares the same principle that the specific endopelvic fascia separates the hypogastric nerve and pelvic plexus from the organ to be resected radically, like the adjacent pelvic surgery such as TaTME and radical prostatectomy.

This study included consecutive cases of women with clinical International Federation of Obstetrics and Gynecologists (FIGO) IA1–IIA cervical cancer who underwent pneumovaginoscopy-assisted radical hysterectomy (PVRH) between 2013 and 2022 at the Kansai Medical University Hospital and Kobe City Medical Center General Hospital (n=59). The vaginal procedures were performed by PV. Abdominal procedures were performed with open (n=38) or laparoscopic surgery (n=21). All PVRH procedures were performed by the same experienced board-certified surgeon incorporating specific measures described below. Patients were offered a choice of treatment, between surgery (open or laparoscopic) or radiation, after preoperative counseling. Type B2 radical hysterectomy according to the Kyoto classification (type II nerve-sparing radical hysterectomy on Viper’s classification) was performed on all patients with pelvic lymphadenectomy, excluding the specific measures described below [ 3 4 ]. This study was approved by the Institutional Review Boards of both hospitals. All patients provided written informed consent for surgery. The platform device (Gel-point path ® ) was inserted and fix-sutured at the inlet of the vagina. To create a vaginal cuff, the tumor margins were first determined using the Schiller test (visual inspection). An incision line was made 2 cm from the tumor margins (green dotted line). Pitressin ® (vasopressin 20 units/A) 1A diluted in saline 100-fold was injected along the incision line to minimize bleeding. The vagina was then incised circumferentially with a monopolar electrocautery device ( Figs. 1A and 2A ). The vagina on the cervix was closed (blue line) with a double layer of continuous sutures using V-Loc ® barbed suture ( Fig. 2B ). The vaginal canal was then thoroughly irrigated with 500 mL of saline. (A-C) Sagittal view. (D) Pneumovaginoscopic view. (E) Anatomical schema around PCEPF. Axial view. Paracolpium is surrounded by PCEPF. Uterine vessels and nerves are located outside of PCEPF. (F) Abdominal surface of PCEPF (abdominal view). PCEPF, paracervical endopelvic fascia, PV, pneumovaginoscopy. (A-D) PV view. (A) Vaginal circumcision, (B) closure of vaginal cuff, (C) resection of paracolpium, and (D) separated PCEPF and vaginal cuff. (E-H) Separated PCEPF, ureter, uterine vessels, and ligaments (laparoscopic view). (I) Resection of PCEPF (laparoscopic view). (J) After completion (laparoscopic view). (K) PCEPF was demonstrated during another laparoscopic hysterectomy (laparoscopic view). PCEPF, paracervical endopelvic fascia; PV, pneumovaginoscopy. The paracolpium, including the vaginal vessels, was dissected (green dotted line) from the surrounding connective tissues with a vessel sealer ( Fig. 1B ). Some penetrating vessels from the paracolpium to the pelvic wall were also dissected (red and blue vessels) ( Fig. 2C ). The vaginal wall was dissected almost bluntly with positive gas pressure (yellow arrows) to reach the dorsal aspect of the attachments of the vesicovaginal septum, recto-uterine septum, and lower uterine ligaments, including the vesicouterine, cardinal, and sacrouterine ligaments. At this stage, these septa and ligaments were observed as one continuous membrane surrounding the cervix, i.e., the paracervical endopelvic fascia (PCEPF) ( Figs. 1D and 2D ). Sheets of gauze with isodine were packed into the vaginal cavity to await convergence with the abdominal approach. The abdominal procedures were open or laparoscopic and performed simultaneously with the PV (2-team surgery). First, conventional pelvic lymphadenectomy was performed, and the upper uterine ligaments (round, infundibulopelvic, and broad ligaments) were dissected. The bladder, rectum, and ureters were also separated from the cervix, and the uterine artery was dissected ( Fig. 2E ). PV light was visible through the PCEPF ( Fig. 2D-H ), and the lower uterine ligaments (observed as the thick portion of the PCEPF) and uterine vessels (including the cervicovesical vessels, superficial and deep uterine veins) were observed on the abdominal surface of the PCEPF ( Figs. 1F , 2E-H ). The uterine vessels were isolated and dissected, usually proximal to the cervix ( Fig. 1C and D ). The deep uterine veins were usually separated and dissected at this time but were sometimes dissected later with the PCEPF. Identifying and separating the hypogastric nerves, pelvic plexus, and bladder nerves were unnecessary ( Fig. 1C and E ). The PCEPF, including the lower uterine ligaments, was dissected from the cervix with a safety margin of approximately 1 cm with an abdominal monopolar electrocautery device or vessel sealers ( Figs. 1C , 2I , and 2J ) under PV observation ( Fig. 1D ). The resected surgical specimen showed that the vagina, cervix, and uterus were packaged within the continuous fascia as a single mass, with the vessels and surrounding PCEPF ( Fig. 3 , Figs. S1 and S2 ). The resected vagina, cervix, and uterine were packed within the continuous fascia as a single mass with paracolpium, parametrium, uterine vessels, and surrounding PCEPF. (A-C) Before specimen dissection. (A) Caudal anterior, (B) anterior, and (C) posterior views. (D-F) After specimen dissection. (D) Caudal anterior, (E) anterior, and (F) posterior views. PCEPF, paracervical endopelvic fascia. Postoperatively, patients considered at risk of recurrence (i.e., with an intermediate risk factor of LVSI, tumor >4 cm in diameter) received adjuvant chemotherapy as described in the guidelines of the Japan Society of Gynecologic Oncology [ 11 ]. Patients with 1 or 2 metastatic lymph nodes were also administered adjuvant chemotherapy. Patients with more risk factors for recurrence received concurrent chemoradiotherapy (CCRT). Upon completion of treatment, patients underwent follow-up examinations every 3 months for post-treatment surveillance, as well as computed tomography at 12-month intervals to screen for tumor recurrence. In cases of subjective symptoms or clinical signs of tumor recurrence, a comprehensive diagnostic workup was initiated. Complications were defined as any event during or after surgery that required another surgical procedure, interventional radiotherapy, or rehabilitation therapy. Normal bladder function was defined as a post-void residual urine volume of <50 mL. Disease-free survival (DFS) was defined as the time between the surgery and the time of initial recurrence or death from cervical cancer, and overall survival (OS) was defined as the time between the surgery and the time of death from any cause. Patients known to be disease-free or alive at their last contact date were censored. Study variables are shown as median (interquartile range) or mean (standard deviation) values. DFS and OS models were constructed using the Kaplan–Meier method. StatFlex ver. 7 (Artech Co., Ltd., Tokyo, Japan). was used in statistical analyses.