This study was conducted prospectively in patients with cervical cancer the International Federation of Gynecology and Obstetrics (FIGO) stage IB1-IIA. Exclusion criteria included neuroendocrine histology, pathologically proven distant metastasis, history of psychiatric disease, preoperative urinary dysfunction, and another coexisting malignancy. Patients with an adequate follow-up duration of more than one year after the surgery were included in the analysis.

Between March 2003 and November 2005, 94 consecutive patients with cervical cancer FIGO stage IB1-IIA were enrolled and 92 patients were randomly assigned to surgical treatment with CRH or NSRH after excluding panic disorder (n=1) and pathologically confirmed supraclavicular lymph node (LN) metastasis (n=1). After random assignment, one patient declined to get included in the CRH group, and one patient with intraoperatively confirmed multiple metastatic LNs was converted to chemoradiation followed by pelvic and para-aortic LND in NSRH group. Two patients in the CRH group who had confirmed neuroendocrine histology in the surgical specimen were excluded from the study, and treated by systemic chemotherapy. One patient in each group was lost to follow-up within one year after the surgery. Eighty-six patients (40 in the CRH group and 46 in the NSRH group) were included in the final analysis (Fig. 1). This prospective study was approved by the Institutional Review Board of the National Cancer Center of Korea.

All operations were performed by either or both surgeons (JWR and SYP) who were skilled in CRH and already had experience with NSRH for one year before this study in the pilot trial. Laparotomy through hypogastric midline incision and left circumcision of the umbilicus was performed for the procedure. A schematic three-dimensional (3D)-illustration of the nervous system of the pelvis and key steps in the surgical procedure are presented in Fig. 2. The key for the nerve-sparing during RH included four steps. The first step was to identify the superior hypogastric plexus (SHP) at the aortic bifurcation and mobilize it from the sacral promontary, which would be the starting point of the nerve-sparing procedure (Fig. 2B). The second step was, after entering the retroperitoneal pelvic space and completion of pelvic LND, the hypogastric nerves (HN) originating from the SHP and the upper part of the inferior hypogastric plexus (IHP) were exposed along the lateral border of the mesorectum (Fig. 2C). The third step was for the IHP (Fig. 2D). The vascular part of the paracervix was ligated at the level where uterine vessels branch from the internal iliac vessels. After incision of the rectouterine peritoneal fold, separation of the anterior mesorectum from the proximal vagina and lateral mobilization of the IHP, which was HN fused with the pelvic splanchnic nerves from S2 to S4, were performed before transaction of the uterosacral ligament. After an anterior incision between the bladder and the cervix, the medial aspect of the distal ureter was exposed by separation of the uterine vessels, lymphatics, and a few small LNs from the bladder mesentery. For complete removal of the paracervical tissues, liposuction was performed with Hercules aspirator (Wells Johnson Co., Tucson, AZ, USA) below the vascular part of the paracervix.

The last part of nerve-sparing was for the bladder branch of the IHP (Fig. 2E). The vesicovaginal venous connection and anterior condensed ligamentous portion crossing the ureters were ligated and transected, and then the vagina at a level 2 to 3 cm distal to the tumor was clamped while saving the posterior part of the vesicouterine ligament which contained the bladder branch of the IHP.

After the insertion of two Jackson-Pratt closed suction drains, suprapubic cystostomy was performed. Bladder emptying with suprapubic catheterization was performed for one week, and then self-voiding was attempted after clamping of the suprapubic catheter. Residual urine (RU) volume was recorded by declamping the catheter immediately after the self-voiding. The suprapubic catheter was removed when the RU measured less than 50 mL at least two times in a low. The duration of bladder catheterization after the surgery was recorded and compared.

For the quantification of nerve fibers retained in the resected paracervix, longitudinal paracervical tissues were marked at 2 cm lateral to the isthmus of the resected uterus at the operating theater (Supplementary Fig. 1). Microscopic section of full length of the labeled paracervix was stained with a general nerve marker, S-100 protein, in order to quantify the nerve fibers. Nerve fibers larger than 100 µm on both sides were counted by the pathologist (SL) without any clinical information and surgical protocol.

Postoperative external radiotherapy (RT) was employed when the final pathologic report revealed LN metastasis or histologically confirmed parametrial invasion, and intracavitary RT was added when the vaginal resection margin was involved by tumor.

Urodynamic studies (UDSs) were carried out before the operation and at 1, 3, and 12 months after the operation. UDS were carried out using the Duet MultiP (Medtronic Inc., Minneapolis, MN, USA) and air-charged urethral and abdominal sensors. All procedures were performed in a sitting erect position on a toilet seat with room-temperature sterile water. The filling rate was 30 mL/min unless this provoked urinary urgency. Abdominal leak point pressures were measured at 200 cm³, with the urethral catheter in place, unless otherwise specified. The abdominal pressure transducer was placed rectally. Provocative maneuvers, including water stimulation and cough, were used in an effort to provoke detrusor overactivity. Urethral pressure profile was performed using a mechanical puller at a rate of 1 mm/sec. Both static and dynamic profiles were performed at cystometric capacity. Recorded urodynamic parameters in addition to the urodynamic tracings were from the uroflowmetry (voided volume, postvoid RU, maximal flow rate, voiding time), cystometrogram (first sensation, first desire to void, capacity, filling rate, detrusor overactivity, urodynamic stress incontinence, abdominal leak point pressure at 200 cm³), and urethral pressure profile. One urologist (JSC) reviewed each case without any clinical information.

Subjective bladder dysfunction was evaluated with a standard questionnaire of the Korean version of International Prostate Symptom Score (IPSS). The scoring system for voiding function described above was based on a 0 to 5 scale, as follows; 0, not at all; 1, less than one time in five; 2, less than half the time; 3, about half the time; 4, more than half the time; and 5, almost always. Questionnaires were analyzed and divided into emptying, frequency, intermittency, urgency, weak stream, straining, and nocturia. Total IPSS score was calculated by adding the score of each item. It can be expected that the greater is the increase in the IPSS, the lesser is the urinary function.

All enrolled patients were followed up every 3 months for the first 2 years after the surgery and every 6 months thereafter with a physical gynecological examination, vaginal cytology, chest X-ray, tumor markers (squamous cell carcinoma antigen and carcinoembryonic antigen), and other imaging modality including computed tomography (CT), magnetic resonance imaging, positron emission tomography (PET), and PET/CT if indicated.

Disease-free survival (DFS) and overall survival (OS) were calculated as time, in months, from the date of the primary surgery to the time of confirmed recurrence of cervical cancer and death, respectively. DFS and OS were compared between the two groups by survival analysis with a long-term follow-up.

Continuous variables were assessed for normal distribution (Kolmogorov-Smirnov test) and expressed as appropriate (mean with SD or median with range). Categorical variables were evaluated with the use of the Fisher exact test. For paired data, such as postoperative changes in UDS or IPSS score compared with preoperative baseline, Wilcoxon singed rank test was used for analysis. A p<0.05 was considered statistically significant. DFS and OS were evaluated by Kaplan-Meier analysis. IBM SPSS ver. 20.0 (IBM Co., Armonk, NY, USA) was used for all statistical analyses.

The detailed clinical and pathologic characteristics of the enrolled patients and the surgical outcomes in both groups are detailed in Table 1. The general characteristics of the patients, including age, BMI, FIGO stage, histologic distribution, tumor size, operating time, estimated blood loss, adjuvant RT, and LN metastasis were not different in the two groups. With respect to the evaluation of radicality of NSRH, surgical outcomes such as the width and length of the resected paracervix, length of safety margin of the resected vagina, and the number of harvested LNs were also similar between the two groups. Average count of resected nerve bundles stained with S-100 in the removed paracervix was significantly lower in the NSRH (12; range, 6 to 21) compared to CRH (30; range, 17 to 45; p<0.001).

The median duration before the postvoid RU volume became less than 50 mL was 11 days (range, 7 to 26 days) among the patients in the NSRH group and was 18 days (range, 10 to 85 days) among the patients in the CRH group (p<0.001). No significant difference in the preoperative data of UDS was found between the two groups (p>0.05). Postoperative changes in UDS compared with preoperative data are shown in Table 2. At postoperative 1 month, almost all parameters of UDS were significantly changed in the CRH group, and several parameters including maximal flow rate, voided volume, RU, compliance, and maximal urethral pressure were also changed in the NSRH group. All changed parameters of USD in the NSRH group were recovered at postoperative 3 months examination, while changes in many parameters of USD in the CRH group were maintained 3 months later. One year after the operation, changes of bladder compliance and RU were still persistent in the CRH group (Fig. 3). For the evaluation of subjective urinary symptoms, postoperative IPSS scores were compared with preoperative IPSS scores (Table 3).

There were changes in every item for urinary symptoms, emptying, frequency, intermittency, urgency, weak stream, straining, and nocturia, after CRH and these changes were persistent for more than a year. On the contrary, there were no changes in all urinary symptoms except for intermittency in the NSRH group after a year, although four items for frequency, intermittency, weak stream, and straining indicated deterioration immediately after the operation.

The median duration of follow-up was 101 months (range, 13 to 137 months) in the CRH group and 111 months (range, 13 to 139 months) in the NSRH group. Ten-year DFS of the patients in the CRH and NSRH groups was 94.9% and 92.4%, respectively (Fig. 4). During the follow-up, two patients in the CRH group and three patients in the NSRH group developed recurrence. Three out of the five patients developed recurrence within 2 years, and one patient was diagnosed with pulmonary metastasis at 7-year follow-up. Four out of the five patients had systemic metastatic diseases, and only one patient was diagnosed para-aortic LN metastasis and was treated by laparoscopic resection followed by RT. One patient in the NSRH group was died with systemic metastatic lesions at 21 months after the initial diagnosis. The clinicopathologic features of the recurred cases are summarized in Supplementary Table 1. The 10-year DFS and OS were not different between the two groups.