Gynecologic Oncology 137 (2015) 291–298

Contents lists available at ScienceDirect

Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno

A prospective study examining the incidence of asymptomatic and symptomatic lymphoceles following lymphadenectomy in patients with gynecological cancer Michal Zikan a,⁎, Daniela Fischerova a, Ivana Pinkavova a, Jiri Slama a, Vit Weinberger b, Ladislav Dusek c, David Cibula a a Gynecological Oncology Center, Department of Obstetrics and Gynecology, Charles University in Prague, First Faculty of Medicine and General University Hospital, Apolinarska 18, Prague 128 00, Czech Republic b Faculty of Medicine, Department of Obstetrics and Gynecology, Masaryk University, Obilni trh 526/11, Brno 602 00, Czech Republic c Institute of Biostatistics and Analyses, Masaryk University, Kamenice 126/3, Brno 625 00, Czech Republic

H I G H L I G H T S • Lymphocele develops in one fifth of patients after lymphadenectomy. • Symptomatic lymphocele is a rare event affecting 5–6% of patients. • Risk factors for lymphocele development are not preventable.

a r t i c l e

i n f o

Article history: Received 26 October 2014 Accepted 16 February 2015 Available online 24 February 2015 Keywords: Lymphocele Pelvic lymphadenectomy Paraaortic lymphadenectomy Radical hysterectomy

a b s t r a c t Objective. To identify the incidence of asymptomatic and symptomatic (i.e., causing pain, hydronephrosis, venous thrombosis, acute lymphedema of the lower or urinary urgency) lymphoceles, as well as risk factors for their development, through a prospective study of patients undergoing sole pelvic or combined pelvic and paraaortic lymphadenectomy for gynecological cancer. Methods. Patients with endometrial, ovarian or cervical cancer scheduled for sole pelvic or combined pelvic and paraaortic lymphadenectomy as a primary surgical treatment or salvage surgery for recurrence were enrolled at single institution from February 2006 to November 2010 and prospectively followed up with ultrasound. Results. Of 800 patients who underwent sole pelvic or combined pelvic and paraaortic lymphadenectomy for gynecological cancer, the overall incidence of lymphoceles was 20.2%, with symptomatic lymphoceles occurring in 5.8% of all patients. Lymphoceles are predominantly located on the left pelvic side wall. Lymphadenectomy in ovarian cancer, a higher number of lymph nodes obtained (N 27), and radical hysterectomy in cervical cancer were found to be independent risk factors for the development of symptomatic lymphoceles. Conclusions. The overall incidence of lymphocele development after lymphadenectomy for gynecological cancer remains high. However, the majority of lymphoceles are only incidental finding without clinical impact. A symptomatic lymphocele is an uncommon event, occurring in only 5.8% of patients. Symptomatic lymphoceles tend to develop earlier than asymptomatic. Although such risk factors are hard to avoid, patients known to be at an increased risk of developing symptomatic lymphoceles can be counseled appropriately and followed up for specific symptoms relating to lymphocele development. © 2015 Elsevier Inc. All rights reserved.

Introduction Lymphocele formation is a complication that occurs following lymphadenectomy due to gynecological or urological malignancy, or after renal transplantation [1–3]. In 1955, Mori published a case series describing the occurrence of 68 lymphoceles following radical ⁎ Corresponding author. Fax: +420 224 967 452. E-mail address: [email protected] (M. Zikan).

http://dx.doi.org/10.1016/j.ygyno.2015.02.016 0090-8258/© 2015 Elsevier Inc. All rights reserved.

hysterectomy and lymphadenectomy for cervical cancer [4]. In 1961, Ferguson and MacClure [5] confirmed that lymphoceles are a complication of the lymphatic system, when a contrast agent injected into a lymphatic vessel penetrated into the lymphocele. The majority of lymphoceles are asymptomatic and are often an incidental finding during postoperative or routine follow-up. Reports of the incidence of asymptomatic lymphoceles following oncogynecological procedures involving lymphadenectomy range from 1% to 58% [4,6–8].

292

M. Zikan et al. / Gynecologic Oncology 137 (2015) 291–298

If symptomatic, lymphocele may result in significant postoperative morbidity and may even delay further cancer treatment. A symptomatic lymphocele may compress adjacent structures (ureters, urinary bladder, rectum or large vessels) and consequently cause pain, hydronephrosis, urinary urgency or thrombosis. The most serious complication of lymphocele is infection [8–13]. An injury to the lymphatic vessels is the key causative factor in the formation of a lymphocele. Recent research has looked to identify potential risk factors which may increase the risk of lymphocele formation after lymphadenectomy [14]. Surgical approach (laparotomy vs. laparoscopy), number of lymph nodes harvested, lymph nodes status, type of cancer, body mass index (BMI), preoperative or adjuvant radiotherapy and chemotherapy are factors discussed as potential risk or protective factors for lymphocele development. The aim of this study was to analyze the incidence of symptomatic and asymptomatic lymphoceles and assess the risk and protective factors for lymphocele formation in the largest cohort of patients that has been followed prospectively after pelvic and/or paraaortic lymphadenectomy for gynecological malignancy, to date.

Scheduled to sole pelvic or combined pelvic and paraaortic lymphadenectomy N=921

Lymphadenectomy performed according to protocol

Drop-out at follow-up N=32

At least one follow-up visit after surgery N=800

3-month follow-up N=800

Study design and patient population

Surgical procedures and techniques Performance of both pelvic and paraaortic lymphadenectomy was precisely standardized. Lymphadenectomy performance was not changed during the study period and was identical for all cancer types [15]. Procedures were done either by laparotomy or laparoscopy. Surgery was undertaken by one of four experienced gynecological oncologists. Pelvic lymphadenectomy started with opening of retroperitoneum parallel to the external and common iliac vessels above the psoas muscle. Lymphatic tissue was removed from the external iliac, common iliac, obturator and interiliac regions by monopolar and bipolar coagulation from the vena circumflexa ilium profunda (caudal limit) up to the aortic bifurcation (cranial limit). Lymphatic tissue from the presacral region was removed separately as a part of paraaortic lymphadenectomy or in cases of cervical cancer as a part of pelvic lymphadenectomy.

Lymphadenectomy not performed according to protocol N=37

N=832

Methods

Patients with endometrial, ovarian or cervical cancer who were scheduled for sole pelvic or combined pelvic and paraaortic lymphadenectomy as a primary surgical treatment or salvage surgery for recurrence were enrolled at one institution from February 2006 to November 2010. Enrollment continued until 800 patients were prospectively enrolled and reached at least one follow-up visit. Patients were excluded from the final analysis set if they were not operated on as scheduled, if either pelvic or paraaortic lymphadenectomy was not performed according to the protocol, or if they did not attend at least one follow-up visit (Fig. 1). All patients who underwent lymphadenectomy as scheduled and who were not excluded were scheduled for regular follow-up visits at 3-month intervals (± 1 month) for 2 years as per the institution's follow-up protocol (according our institution´s guidelines, every patient after surgical staging undergoes ultrasound examination every 3 months) or as required in case of presentation of symptoms. All symptoms were carefully registered at each follow-up visit. If symptoms occurred in any patient with a lymphocele and the nature of their symptoms could be attributed to its presence, the lymphocele was considered to be symptomatic. The following symptoms were considered potentially related: pain located to the site of lymphocele, hydronephrosis, urinary urgency, venous thrombosis and acute lymphedema of the lower extremity. This study has been approved by a local ethical committee and all patients gave their informed consent.

Surgery abandoned N=52

6-month follow-up N=793 9-month follow-up N=726 12-month follow-up N=701 >12-month follow-up N=690 Fig. 1. Study design and patient flow.

For paraaortic lymphadenectomy, the procedure commences with the opening of the retroperitoneum along the root of the mesentery. The aorta and vena cava were identified as well as the ureters, ovarian veins, inferior mesenteric artery and renal veins. Lymphatic tissue was removed from the paracaval, interaortocaval, paraaortic and presacral regions. The caudal limit of dissection was the bifurcation of the aorta, with the cranial limit being the level of both renal veins. Identical extend and performance was used for laparoscopic approach. In all laparoscopic cases, transperitoneal access was used. The peritoneum was left open after surgery and a suction drain was inserted into the pelvis through the abdominal wall in all (either laparotomic or laparoscopic) cases and was removed third postoperative day. A single dose of prophylactic antibiotics was administered intraoperatively.

Imaging Transabdominal and transvaginal or transrectal ultrasound was undertaken according to institutional follow-up guidelines. Examinations were performed by one of three gynecological oncologists experienced in the field of ultrasound diagnostics. GE Logiq 9 or GE 8 instruments in B-mode and power Doppler mode were used. Examination began with transabdominal evaluation of the parenchymatous organs. The superior mesenteric artery was followed into the mesenteric branches in order to check the root of the mesentery. The retroperitoneum was screened in longitudinal and transversal sections from the coeliac trunk to the bifurcation of aorta. Subsequently, both groin and iliac vessels were examined upward from the femoral

M. Zikan et al. / Gynecologic Oncology 137 (2015) 291–298

vessels to the aortic bifurcation. Examination concluded with transvaginal or transrectal scan [16]. If any lesion was found, it was evaluated according to IOTA terms and definitions [17]. A lymphocele was defined as a uni- or multilocular, oval, round or hourglass shaped cystic structure with a thick wall and fluid content of varying echogenicity, which might contain thin septae and debris (Fig. S1). For all identified lymphoceles, the following parameters were recorded: size in three dimensions (craniocaudal, anterioposterior, transverse), location (paraaortic, common iliac, external iliac, obturator fossa), shape (oval, round, hourglass), echogenicity (anechoic, ground-glass, low-level, hemorrhagic, mixed) and the presence of internal structures (septae, debris). In cases where there was doubt relating to the nature of the lesion, computed tomography (CT) was carried out. In cases of persisting uncertainty, drainage with cytological evaluation of content (if the lesion was cystic) or a Tru-Cut biopsy (if a solid component was present) was taken under ultrasound guidance to exclude recurrence of disease.

Statistics Standard descriptive statistics were applied to describe primary data; continuous variables were expressed as arithmetic mean and standard deviations (SD), supplemented by median and minimum– maximum ranges; absolute and relative frequencies of categories were used for the description of categorical variables. Statistical significance of differences between and among groups of patients was tested for continuous variables by the Mann–Whitney U test or the Kruskal– Wallis test; a maximum likelihood chi-square test was used for categorical variables. Characteristics predicting the occurrence of different types of lymphocele were identified using logistic regression analysis. The predictive power of identified protective and risk factors was quantified by odds ratios (ORs) with corresponding 95% confidence intervals. Both univariate and multivariate-adjusted ORs were estimated. A value of P b 0.05 was considered as the level of statistical significance in all analyses. Analyses were performed using SPSS 21 (IBM Corporation, 2012).

293

Results From February 2006 to November 2010, 921 patients were enrolled to follow the study protocol as they were scheduled for pelvic and/or paraaortic lymphadenectomy by either laparotomy or laparoscopy. Lymphadenectomy was performed according to the study protocol in 832 patients, with 800 patients having at least one follow-up visit (Fig. 1). Of the 800 patients who were followed up according to the study protocol, ultrasound revealed thick wall cystic retroperitoneal structures identified to be lymphoceles in 161 patients (20.1%). In total, 227 lymphoceles were found in these 161 individuals. A total of 115 patients out of 161 identified to have lymphoceles were asymptomatic (with no symptoms related to the lymphoceles); 46 patients (28.6% of all cases with lymphoceles and 5.8% of all patients) were symptomatic (pain, hydronephrosis, urinary urgency, venous thrombosis, sepsis, acute lymphedema) (Fig. 2). The median duration of follow-up was 37.2 months (range: 0.5–72.9). Median time to diagnosis of either asymptomatic or symptomatic lymphoceles was 4.8 months (range: 0.5–14.5). Symptomatic lymphoceles tended to appear earlier than asymptomatic ones, at 3.7 (range: 0.5–7.7) months vs. 5.0 (range: 0.4–16.9) months, respectively. No lymphocele was misinterpreted as a recurrence and vice versa. The mean size (largest diameter) of lymphoceles was 47.8 mm (10–180 mm). There was a significant difference between asymptomatic lymphocele (mean size 40.2 mm; 10–180 mm) and symptomatic lymphocele (mean size 67.3 mm; 26–168 mm) (P b 0.001). Of the 227 lymphoceles identified, 220 (96.9%) were localized to the pelvis (common iliac, external iliac or obturator fossa regions), with only 7 (3.1%) being found in the paraaortic region. The most frequent site of localization was along the external iliac vessels (136/221 lymphoceles [61.5%]). Lymphoceles were localized predominantly on the left side of the pelvis (138/220 lymphoceles [63.2%]) with no significant difference in the site of localization being observed between symptomatic and asymptomatic patients. Full patient characteristics and differences between groups with and without lymphoceles (both symptomatic and asymptomatic) are given in Table 1. The median age of patients at surgery was 55 years

Pelvic +/– paraaortic lymphadenectomy for gynecological cancer N=800

Ultrasound every 3 months or if symptomatic

No lymphocele N=639 (79.87%)

Lymphocele N=161 (20.13%)

Asymptomatic lymphocele N=115 (14.38%)

Symptomatic lymphocele N=46 (5.75%)

Inflammation N=14 (1.75%)

Other symptoms N=32 (4.00%)

Fig. 2. Flow chart of principal results. Data described by absolute counts (and percentages of all patients).

294

M. Zikan et al. / Gynecologic Oncology 137 (2015) 291–298

Table 1 Patient characteristics and comparison of patients with and without lymphoceles. Characteristic

All patients1 N = 800

(1) Patients without lymphocele1 N = 639

(2) Patients with lymphocele1 N = 161

Age at surgery (years)

55.0 (19.0; 87.0) 53.9 (13.6) 26.4 (16.2; 60.2) 27.7 (6.6) 70.8 (44.0; 166.0) 75.3 (18.0)

56.0 (19.0; 87.0) 54.4 (13.6) 26.6 (16.2; 60.2) 28.0 (6.8) 72.0 (45.0; 166.0) 76.2 (18.5)

53.0 (20.0; 82.0) 52.0 (13.5) 25.0 (17.1; 47.3) 26.4 (5.7) 70.0 (44.0; 131.0) 71.6 (15.5)

Type of cancer Endometrial cancer Cervical cancer Ovarian cancer Vulvar cancer

307 (38.4%) 292 (36.5%) 173 (21.6%) 28 (3.5%)

264 (41.3%) 245 (38.3%) 107 (16.7%) 23 (3.6%)

43 (26.7%) 47 (29.2%) 66 (41.0%) 5 (3.0%)

0.001 0.035 b0.001 0.987

Type of lymphadenectomy Sole pelvic Combined pelvic and paraaortic Radical hysterectomy

469 (58.6%) 331 (41.4%) 220 (27.5%)

410 (64.2%) 229 (35.8%) 174 (27.2%)

59 (36.6%) 102 (63.4%) 46 (28.6%)

b0.001

660 (82.5%) 140 (17.5%) 37 (0; 130) 38.4 (21.2) 125 (15.7%) 3 (1; 63) 5.5 (9.0) 204 (25.5%) 297 (37.1%)

512 (80.1%) 127 (19.9%) 35 (0; 130) 36.4 (20.8) 89 (14.0%) 3 (1; 56) 5.2 (7.9) 178 (27.9%) 218 (34.1%)

148 (91.9%) 13 (8.1%) 45 (2; 113) 46.6 (20.9) 36 (22.4%) 2 (1; 63) 6.3 (11.3) 26 (16.1%) 79 (49.1%)

b0.001

BMI (kg/m2) Weight (kg)

Approach Laparotomy Laparoscopy Total no. of nodes obtained Patients with positive nodes Number of positive nodes Postoperative radiotherapy Postoperative chemotherapy

P 0.057 0.005 0.005

0.767

b0.001 0.011 0.733 0.002 b0.001

1 Categorical data are described with absolute numbers and the percentage of patients in each given category; continuous variables are described by median (min; max) and mean (SD); follow-up is described using median (5th; 95th percentile). 2 Overall statistical significance of differences between groups was tested by ML-χ2 or Fisher's exact test for categorical data and by Mann–Whitney U test for continuous parameters. Statistically significant results are marked in bold.

(range: 19–87). Median weight and BMI were 70.8 kg (range: 44.0–166.0) and 26.4 kg/m2 (range: 16.2–60.2), respectively. The most frequent type of cancer was endometrial cancer (307 patients; 38.4%) followed by cervical (292 patients; 36.5%), ovarian (173 patients; 21.6%) and other rare cancers (28 patients; 3.5%). Pelvic lymphadenectomy was performed in 469 patients, with pelvic and paraaortic lymphadenectomy being performed in 331 patients. The median number of lymph nodes obtained from pelvic lymphadenectomy was 31 (range: 2–77), from combined lymphadenectomy 52 (range: 2–130). Lymph nodes were positive in 125 patients (15.7%). A laparoscopic approach was chosen in 140 patients (17.5%) with laparotomy being undertaken in 660 patients (82.5%). Radical hysterectomy was part of the procedure in 220 patients (27.5%), with all but two which being carried out by laparotomy. Preoperative radiotherapy, either for the management of recurrent disease or for another medical condition, was rare (13 patients; 1.6%). Neoadjuvant chemotherapy for ovarian, cervical or, rarely, endometrial cancer was administered in 63 patients (7.9%). A total of 204 patients (25.5%) received postoperative radiotherapy, with postoperative chemotherapy being administered in 297 patients (37.1%). Between groups of patients with and without lymphocele, no significant difference was observed in age at surgery (P = 0.057); however, weight and BMI were significantly lower in the lymphocele group (P = 0.005). Lymphoceles were found more frequently in patients with ovarian cancer (P b 0.001) compared with those with endometrial (P = 0.001) and cervical cancer (P = 0.035). In addition, there was a significant difference in the type of cancer that gave rise to symptomatic and asymptomatic lymphoceles, with the proportion of cervical cancer patients being higher in the group with symptomatic lymphoceles (Table 2). The frequency of lymphoceles was significantly higher in patients who underwent combined pelvic and paraaortic lymphadenectomy compared to those who had pelvic lymphadenectomy only (P b 0.001).

Patients who developed lymphocele had a significantly higher number of lymph nodes harvested compared with those who did not develop lymphocele (35 vs. 45; P b 0.001), whereas there was no difference in the average number of lymph nodes harvested between patients with symptomatic and asymptomatic lymphoceles. The positivity of lymph nodes was more frequent in the asymptomatic lymphocele group only (Table 2). Grossly positive bulky lymph nodes were significantly more often in lymphocele group (P = 0.011); however, there were no difference in bulky lymph node frequency between symptomatic and asymptomatic lymphocele group (P = 0.289). Laparotomy increased the incidence of both symptomatic and asymptomatic lymphocele compared to laparoscopy (P b 0.001). Postoperative chemotherapy was more frequent in asymptomatic lymphocele group (P b 0.001), while postoperative radiotherapy protected from any lymphocele formation (P b 0.009) (Table 2). The distribution of stage of the disease (irrespective of type of cancer) did not differ between patients with and without lymphocele. In a multivariate analysis of risk factors for formation of any lymphocele (asymptomatic or symptomatic), four monitored variables were identified as independent risk factors (Table 3): weight below 75 kg (OR = 1.766 [95% CI: 1.170; 2.667]; P = 0.007), combined lymphadenectomy (OR = 2.034 [95% CI: 1.265; 3.270]; P = 0.003), total number of lymph nodes obtained in endometrial cancer patients (OR = 1.848 [95% CI: 1.056; 3.504]; P = 0.038) and in ovarian cancer patients (OR = 3.137 [95% CI: 1.882; 5.229]; P b 0.001) and radical hysterectomy in cervical cancer patients with positive lymph nodes (OR = 1.564 [95% CI: 1.071; 2.487]; P = 0.043). With respect to the risk of symptomatic lymphoceles development, a multivariate analysis identified three major risk factors (Table 4): diagnosis with ovarian cancer (OR = 2.310 [95% CI: 1.089; 4.904], P = 0.029), radical hysterectomy in patients with cervical cancer (OR = 2.208 [95% CI: 1.058; 4.606]; P = 0.035) and the total number of lymph nodes obtained being N27 (1.858 [95% CI: 1.021; 3.381]; P = 0.046).

M. Zikan et al. / Gynecologic Oncology 137 (2015) 291–298

295

Table 2 Comparison of patients without, with asymptomatic and with symptomatic lymphoceles. Characteristic

A. Patients without lymphocele1 N = 639

B. Asymptomatic lymphocele1 N = 115

C. Symptomatic lymphocele1 N = 46

Age at surgery (years)

56.0 (19.0; 87.0) 54.4 (13.6) 26.6 (16.2; 60.2) 28.0 (6.8) 72.0 (45.0; 166.0) 76.2 (18.5)

55.0 (20.0; 82.0) 52.4 (13.5) 24.7 (17.1; 43.9) 26.2 (5.7) 70.0 (44.0; 131.0) 71.1 (15.3)

51.5 (27.0; 81.0) 51.0 (13.7) 26.6 (19.4; 47.2) 26.8 (5.5) 70.0 (50.0; 130.0) 72.7 (16.3)

BMI (kg/m2) Weight (kg)

P 0.091 0.136 0.201

Type of cancer Endometrial cancer Cervical cancer Ovarian cancer Other

264 (41.3%)* 245 (38.3%)* 107 (16.7%)# 23 (3.7%)

29 (25.2%)# 30 (26.1%)# 51 (44.3%)* 5 (4.4%)

14 (30.4%) 17 (37.0%)* 15 (32.6%) 0 (0.0%)

0.002 0.037 b0.001 0.658

Type of lymphadenectomy Sole pelvic Combined pelvic and paraaortic Radical hysterectomy

410 (64.2%)* 229 (35.8%)# 174 (27.2%)

37 (32.2%)# 78 (67.8%)* 30 (26.1%)

22 (47.8%) 24 (52.2%) 16 (34.8%)

b0.001

512 (80.1%)# 127 (19.9%)* 35 (0; 130) # 36.4 (20.8) 89 (14.0%)# 3 (1; 56) 5.2 (7.9) 178 (27.9%)* 218 (34.1%)#

105 (91.3%)* 10 (8.7%)# 45 (2; 113)* 46.6 (21.7) 29 (25.2%)* 3 (1; 23) 5.4 (6.2) 17 (14.8%)# 62 (53.9%)*

43 (93.5%)* 3 (6.5%)# 47 (12; 90)* 46.6 (18.9) 7 (15.2%)# 1 (1; 63) 10.1 (23.3) 9 (19.6%) 17 (37.0%)#

Approach Laparotomy Laparoscopy Total no. of nodes obtained Patients with positive nodes Number of positive nodes Postoperative radiotherapy Postoperative chemotherapy

0.506

0.002 b0.001 0.010 0.289 0.009 b0.001

Statistically significant differences within characteristic groups/categories (*highest value; #lowest value). 1 Categorical data are described with absolute numbers and the percentage of patients in each given category; continuous variables are described by median (min; max) and mean (SD); follow-up is described using median (5th; 95th percentile). 2 Overall statistical significance of differences among groups was tested by ML-χ2 for categorical data and by Kruskal–Wallis test for continuous parameters. Statistically significant results are marked in bold.

Discussion In the largest cohort of patients undergoing pelvic and/or paraaortic lymphadenectomy for gynecological cancer to date, we have assessed the frequency of symptomatic and asymptomatic lymphoceles as well as the factors that may influence the formation of lymphocele. In the full study group of 800 patients who underwent lymphadenectomy according to the protocol and had at least one follow-up visit, lymphocele was detected in 161 patients (20.1%). Reporting of the incidence of symptomatic and asymptomatic lymphoceles has varied greatly across studies. Mori [4] reported an average incidence of 48.5% in 1955; however, more recent studies have reported much lower occurrence rates. Ilancheran and Monaghan [18] found the incidence of symptomatic lymphoceles to be 25.3% in 221 cervical cancer patients. Conte et al. [11] noted an incidence of 22% in 36 cervical cancer patients, Benedetti-Panici et al. [8] reported an incidence of 10–25% in a mixed group of 137 gynecological cancer patients and Logmans et al. [19] cited an incidence of 20–32% in 22 gynecological cancer patients. This considerable variability might be due to different methods being used for lymphocele detection (clinical examination, ultrasound, CT). The largest sample of patients (264 individuals) published so far [9] is a retrospective study where lymphoceles were detected through clinical examination, and only if they were considered to be clinically suspicious, than confirmed with imaging. As most lymphoceles are asymptomatic, they are typically only incidentally identified upon routine postoperative imaging. Surgical technique and technologies (e.g., use electrosurgery in lymphadenectomy) are developing rapidly, which makes older reports unsuitable for comparison [4,5]. In addition, some authors have used a different radicality of lymphadenectomy according to tumor type [20]. In this prospective study, the procedure for lymphadenectomy was highly standardized and was not modified throughout the duration of the study period; this can be documented through the consistent numbers of lymph nodes harvested in individual years, as well as the recorded median

numbers of lymph nodes obtained from sole pelvic or combined lymphadenectomy being quite high (31 and 52, respectively). Although lymphoceles are widely considered to be of limited clinical significance, they can cause noticeable clinical symptoms, mostly from their pressure on adjacent structures or organs and which relate to their size and site of occurrence. An additional consideration is the possibility of the content of lymphoceles becoming infected, via hematogenous or lymphatic dissemination, or via local spread. Lymphocele infection is a potentially serious complication, which may result in sepsis. Fortunately, only a minority of lymphoceles (5–34.5%) become symptomatic [8–13]. The broad range of reported incidences may also be due to different definitions of what comprises a symptomatic lymphocele. In some studies, large lymphoceles detected by clinical exam are also considered to be 'symptomatic' [9]. In our study, lymphoceles were considered symptomatic only if they caused clinical symptoms. We found the incidence of symptomatic lymphoceles on the lower limit of the reported range, with 46 symptomatic lymphoceles in a group of 800 patients (5.8%). In this study, the mean time to lymphocele diagnosis was 3.7 months (range: 0.5–7.7) for symptomatic and 5.0 months (range: 0.4–16.9) for asymptomatic patients. Intervals of several weeks to several months from surgery until detection of lymphocele have been reported previously [10]. In a study of the natural history of pelvic lymphoceles, 40% of 108 patients were found to have lymphoceles, as detected by repeated routine ultrasonography; 81% of all cases were identified 2 weeks after surgery, whereas no new lymphoceles were detected more than 6 months after surgery [21]. We have not found any existing data regarding the predominant location of lymphoceles in the literature. In this report, the majority of lymphoceles were localized in the pelvis (220/227; 96.9%), predominantly on the left pelvic side wall (138/220; 63.2%). This observed imbalance may be due to the different anatomy of lymphatic drainage on the different sides of the pelvis [15]. Lymphatic system of the pelvis is reported to be asymmetric, with more (up to 2×) nodes in the left iliac region [22–24].

296

M. Zikan et al. / Gynecologic Oncology 137 (2015) 291–298

Table 3 Analysis of risk factors for the occurrence of lymphoceles in the full data set (N = 800). Univariate

Multivariate—adjusted

Predictor

Risk category

OR (95% CI)

Age BMI Weight Type of cancer

≤60 ≤28 ≤75 Endometrial cancer Cervical carcinoma Ovarian cancer Other Combined pelvic and paraaortic

1.550 (1.053; 2.284) 1.609 (1.111; 2.331) 1.848 (1.263; 2.704) reference 1.178 (0.752; 1.845) 3.787 (2.427; 5.909) 1.335 (0.482; 3.699) 3.095 (2.162; 4.432)

Type of cancer Endometrial cancer Cervical cancer Ovarian cancer Other Radical hysterectomy (without ovarian cancer patients)

Type of lymphadenectomy Combined pelvic and paraaortic Combined pelvic and paraaortic Combined pelvic and paraaortic Combined pelvic and paraaortic Yes

2.613 (1.333; 5.123) 1.836 (0.803; 4.194) 6.072 (1.750; 21.070) 1.250 (0.172; 9.093) 1.948 (1.253; 3.029)

0.005 0.150 0.004 0.826 0.003

Type of cancer Cervical cancer Approach

Radical hysterectomy Yes Laparotomy

4.877 (1.862; 12.776) 2.824 (1.551; 5.142)

0.001 b0.001

Type of cancer Cervical cancer Total no. of nodes

Approach Laparotomy N27

4.792 (1.829; 12.555) 2.997 (1.899; 4.730)

0.001 b0.001

Type of cancer Endometrial cancer Cervical cancer Ovarian cancer Other Positive nodes Postoperative radiotherapy Postoperative chemotherapy

Total no. of nodes N27 N27 N27 N27 Yes No Yes

2.236 (1.055; 5.207) 1.858 (0.947; 3.646) 7.087 (2.056; 24.434) 4.364 (0.421; 45.257) 1.780 (1.154; 2.744) 1.990 (1.264; 3.134) 1.884 (1.327; 2.674)

0.034 0.072 0.002 0.217 0.009 0.003 b0.001

Type of lymphadenectomy

P

OR (95% CI)

P

0.026 0.012 0.002

1.766 (1.170; 2.667)

0.007

0.475 b0.001 0.579 b0.001

2.034 (1.265; 3.270)

0.003

1.848 (1.056; 3.504)

0.038

3.137 (1.882; 5.229)

b0.001

1 ORs and analyses of statistical significance were derived using logistic regression. Statistically significant results are marked in bold.

An injury to the lymphatic vessels is the key causative factor in the formation of a lymphocele. Recent research has focused on examining a number of potential risk factors, which may increase the risk of lymphocele formation following lymphadenectomy [14]. BMI is often quoted as a risk factor. In a retrospective study by Kim et al. [9], a significantly higher BMI was found in the group of patients who develop postoperative lymphoceles (23.94 ± 3.38 vs.

22.52 ± 3.00, P b 0.01). Conversely, in a study of 88 patients following pelvic or paraaortic lymphadenectomy for gynecological malignancy, Achouri et al. [10] did not observe a higher risk of lymphocele formation associated with higher BMI (P = 0.34). Surprisingly, in this study, univariate analysis highlighted that a lower BMI (≤ 28) and weight (≤ 75 kg) were risk factors for the formation of lymphoceles, although they were not risk factors for

Table 4 Analysis of risk factors for the occurrence of symptomatic lymphoceles in the full data set (N = 800). Univariate

Multivariate—adjusted

Predictor

Risk category

OR (95% CI)

P

Age BMI Weight Type of cancer

≤60 ≤28 ≤75 Endometrial cancer Cervical cancer Ovarian cancer Others Combined pelvic and paraaortic Yes

1.645 (0.822; 3.294) 1.502 (0.788; 2.862) 1.445 (0.758; 2.753) Reference 1.294 (0.626; 2.675) 1.987 (1.035; 4.021) 1.588 (0.875; 2.884) 2.065 (1.001; 4.262)

0.160 0.216 0.264 0.487 0.034 0.128 0.049

Laparotomy

3.183 (1.001; 10.409)

0.048

Approach Laparotomy Laparotomy Laparotomy Laparotomy N27 Yes No Yes

0.840 (0.181; 3.909) 8.176 (1.068; 62.606) – – 2.273 (1.044; 4.945) 0.967 (0.423; 2.215) 1.437 (0.681; 3.031) 0.990 (0.534; 1.835)

0.825 0.043 0.039 0.937 0.341 0.975

Type of lymphadenectomy Radical hysterectomy (without ovarian cancer patients) Approach Type of cancer Endometrial cancer Cervical cancer Ovarian cancer Other Total no. of nodes Positive nodes Postoperative radiotherapy Postoperative chemotherapy

1 ORs and analyses of statistical significance were derived using logistic regression. Statistically significant results are marked in bold.

OR (95% CI)

P

2.310 (1.089; 4.904)

0.029

1.858 (1.021; 3.381)

0.046

M. Zikan et al. / Gynecologic Oncology 137 (2015) 291–298

the formation of symptomatic lymphoceles. This trend for lymphocele formation can be hypothetically caused by a possible tendency to be more radical with lymph node dissection in slim patients, which is outside the constraints of the standardization of the procedure. Several studies have shown that among gynecological malignancies, the highest incidence of lymphoceles occurs in patients with cervical cancer, followed by patients with ovarian cancer and patients with endometrial cancer [9]. On the contrary, we observed the highest incidence of both asymptomatic and symptomatic lymphoceles in ovarian cancer patients (66 patients with lymphoceles out of 173 ovarian cancer patients; 38.2%) and a lower incidence in endometrial (43/307; 14.0%) and cervical cancer (47/292; 16.1%). Following multivariate analysis, the diagnosis of ovarian cancer was found to be a significant risk factor for the formation of any lymphoceles and for the formation of symptomatic lymphocele (OR = 3.137 and OR = 2.310, respectively). We can speculate that this difference of incidence between our study and other studies may be due to two factors—performance of lymphadenectomy was identical for all tumors types in our study, while lymph node dissection is commonly undertaken to different extents in ovarian, endometrial and cervical cancer patients. Additionally, complex surgery in advanced ovarian cancer, including peritonectomy, may potentially reduce the ability of the peritoneum to absorb excess lymph. However, when testing this hypothesis, we did not find any difference in lymphocele formation (neither symptomatic or asymptomatic) between a group of 93 patients after standard staging procedure for ovarian cancer (i.e., total hysterectomy and bilateral salpingo-oophorectomy, total omentectomy, appendectomy, pelvic and paraaortic lymphadenectomy)—37 patients with lymphocele (56.1%)—and a group of 80 patients after excessive debulking for advanced ovarian cancer (these surgeries included pelvic peritonectomy, bowel resections and/or diaphragmatic surgery in addition)—29 patients with lymphocele (43.9%); P = 0.642. Hence, the explanation remains unclear. There are only limited data on potential influence of the type of lymphadenectomy (pelvic vs. pelvic and paraaortic lymphadenectomy) on lymphocele development. In the work of Achouri et al. [10], the addition of paraaortic lymphadenectomy did not increase the risk of lymphocele development (OR = 2.4; 95% CI 0.7–8.0). Conversely, we found the combination of pelvic and paraaortic lymphadenectomy to be a significant risk factor for lymphocele development (OR = 2.034, P = 0.003), but not for formation of symptomatic lymphoceles (OR = 1.588, P = 0.128). We speculate that performance of paraaortic lymphadenectomy, together with pelvic lymphadenectomy, can lead to disruption of the accessory lymphatic pathways, worsening the lymphatic drainage from the lower limbs and abdomen, and thus increasing the leakage of lymph in the pelvis. There was no difference in the number of lymph nodes obtained in particular tumor types. The choice of surgical approach, i.e., laparoscopy or laparotomy, was discussed as another factor. The largest report on this variable published to date [25] studied a group of patients operated on for endometrial cancer and found a significantly lower incidence of lymphocele in patients following laparoscopic approach. The lymphocele incidence in laparoscopy patients was found to be 1.4%, i.e., 2 cases. Following laparotomy, lymphoceles developed in 19 cases, i.e., the incidence of lymphocele amounted to 15.4% (OR = 12.42; 95% CI 2.82–54.55; P b 0.0001). Additionally, symptomatic lymphoceles occurred with greater frequency in patients following laparotomy than in those treated laparoscopically (P = 0.028). However, a recent study of 88 patients [10] did not find such a difference; in 31 lymphocele patients, with 11 having a laparoscopy and 20 having a laparotomy (OR = 0.61; 95% CI 0.25–1.52). In our study, the incidence of lymphoceles in the laparotomy group was 148/660 (22.4%) and 13/140 (9.3%) in the laparoscopy group. We found laparotomy to be a risk factor for the occurrence of any lymphoceles as well as symptomatic lymphoceles (OR = 2.824, P b 0.001 and OR = 3.183, P = 0.048, respectively). Preoperative or postoperative radiotherapy as well as adjuvant or neoadjuvant chemotherapy are considered to be risk factors for

297

lymphocele formation [12]. A study by Kim et al. [9] assessing a group of 264 patients (50 lymphoceles) reported a significantly higher incidence of asymptomatic lymphoceles in groups of patients treated with postoperative radiotherapy (29.2% vs. 15.6%; P = 0.01), although data on the proportion of symptomatic lymphoceles were not shown. Achouri et al. [10] described a small cohort of 31 patients with symptomatic lymphoceles and did not observe a higher incidence in the group who had been treated with radiotherapy (OR = 0.25; 95% CI 0.06–0.98). Patients who have received neoadjuvant or postoperative chemotherapy have not been shown to have a higher incidence of lymphoceles [9]; however, data are inconsistent on this topic too. In this study, we found a significantly lower incidence of lymphocele formation in the group of patients who received postoperative radiotherapy (26/204; 12.7%) compared to those who had no postoperative radiotherapy (134/595; 22.5%) (P = 0.002). This observation may reflect the occurrence of an extensive fibrotic reaction of the irradiated tissue, including the lymphatic channels, after radiation. Postoperative chemotherapy was found to be a risk factor (OR = 1.884, P b 0.001) for overall lymphocele formation, but not for symptomatic lymphocele, following multivariate analysis. The number and positivity of lymph nodes removed has also been reported to increase the incidence of lymphoceles. Petru et al. [12] found a higher incidence of lymphocele in cervical cancer patients whose nodes were positive, as opposed to lymph node negative patients (29% vs. 14%, P b 0.02). Kim et al. [9] documented a group of patients with postoperative lymphoceles that had a greater number of nodes resected during surgery than a group that developed no postoperative lymphoceles (26.80 ± 14.82 vs. 22.96 ± 10.18, P = 0.03); other studies, however, have failed to confirm such a relationship. These two risk factors appear to be more prominent in larger cohorts [8,10,26]. In concordance with Petru et al. [12], we found a significant difference in the total number of lymph nodes obtained between the groups without and with lymphocele (a median of 35 vs. 45 lymph nodes, respectively; P b 0.001). In predictive modeling, a total number of lymph nodes N 27 was found to be a risk factor for the formation of any lymphocele (OR = 2.997, P b 0.001) as well as an independent risk factor for the formation of symptomatic lymphocele (OR = 1.858, P = 0.046 in multivariate analysis). Node positivity was found to be a risk factor for the occurrence of any lymphoceles only (OR = 1.780, P = 0.009) following univariate analysis. If signs of inflammation or serious symptoms develop, the lymphocele must be treated. It is always our aim to be non-invasive. Percutaneous evacuation and drainage under ultrasound or CT guidance is a simple method, which may be combined with the administration of sclerotizing agent, most frequently ethanol or betadine. The overall success rate of repeated drainage is 90–100%. Simple aspiration alone has limited treatment effect. Betadine must not be used in patients with documented iodine allergy. An increased incidence of infectious complications of 9–15% has been reported after percutaneous drainage. Our policy is to administer prophylactic antibiotics before and during the procedure. In this study, we used a monopolar and bipolar electrocoagulation for lymph nodes dissection. There are several reports in the literature on new sources of energy or new surgical devices referring 10× lower incidence of postoperative pelvic lymphocele for Ligaclip [27] and lower incidence of lymphorrhoea after axillary dissection for Ultracision [28]. Other studies are needed to validate the role of new surgical devices in reduction of lymphocele incidence. The potential current way, how to decrease lymphocele incidence by decreasing the number of lymph nodes harvested, is to avoid systematic lymphadenectomy if not necessary, especially in stages of malignant disease with low risk of nodal involvement. Sentinel lymph node mapping may in these cases avoid understaging on one side but also avoid side effects associated with systematic lymphadenectomy. Sentinel lymph node dissection is currently used in many institutional for cervical cancer and early-stage endometrial cancer [29,30].

298

M. Zikan et al. / Gynecologic Oncology 137 (2015) 291–298

As only symptomatic lymphocele requires clinical attention and intervention, the incidence of symptomatic lymphoceles represents the most clinically important outcome. However, as any lymphocele, either symptomatic or asymptomatic, may be misinterpreted on imaging as a suspicious recurrence, knowledge of the likely incidence of lymphoceles, their location and key risk factors may help to avoid misdiagnosis. Conclusions In the largest prospective study so far, we showed that the incidence of any lymphocele after pelvic and/or paraaortic lymphadenectomy for gynecological malignancy is quite high (20.1%). However, the majority of lymphoceles are asymptomatic and only found incidentally and without clinical impact. A symptomatic lymphocele is a rare event occurring in only 5.8% of all the patients. Lymphoceles are predominantly (63.2%) located on the left pelvic wall with symptomatic lymphoceles tending to develop earlier than asymptomatic (3.7 vs. 5.0 months). We found that lymphadenectomy for ovarian cancer, higher number of lymph nodes obtained (N27) and radical hysterectomy in cervical cancer patients were independent risk factors for the development of symptomatic lymphocele. Although these factors are hard to avoid, patients known to be at an increased risk of developing symptomatic lymphocele can be counseled appropriately and monitored during their follow-up for any relevant symptoms that might indicate the development of a symptomatic lymphocele. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ygyno.2015.02.016. Conflict of interest None of the authors have any potential conflicting interests to declare.

[9]

[10] [11]

[12]

[13] [14]

[15]

[16] [17]

[18] [19]

[20]

[21]

[22] [23]

Acknowledgments This work was supported by the Internal Grant Agency of the Ministry of Health of the Czech Republic (project no. NT 13070) and by the Charles University in Prague (project nos. UNCE 204024 and PRVOUK P27/LF1/1).

[24] [25]

[26]

References [1] Glass LL, Cockett AT. Lymphoceles: diagnosis and management in urologic patients. Urology 1998;51:135–40. [2] Dodd GD, Rutledge F, Wallace S. Postoperative pelvic lymphocysts. Am J Roentgenol Radium Ther Nucl Med 1970;108:312–23. [3] Braun WE, Banowsky LH, Straffon RA, Nakamoto S, Kiser WS, Popowniak KL, et al. Lymphocytes associated with renal transplantation. Report of 15 cases and review of the literature. Am J Med 1974;57:714–29. [4] Mori N. Clinical and experimental studies on the so-called lymphocyst which develops after radical hysterectomy in cancer of the uterine cervix. J Jpn Obstet Gynecol Soc 1955;2:178–203. [5] Ferguson JH, Maclure JG. Lymphocele following lymphadenectomy. Am J Obstet Gynecol 1961;82:783–92. [6] Gray MJ, Plentl AA, Taylor Jr HC. The lymphocyst: a complication of pelvic lymph node dissections. Am J Obstet Gynecol 1958;75:1059–62. [7] Mann WJ, Vogel F, Patsner B, Chalas E. Management of lymphocysts after radical gynecologic surgery. Gynecol Oncol 1989;33:248–50. [8] Benedetti-Panici P, Maneschi F, Cutillo G, D'Andrea G, di Palumbo VS, Conte M, et al. A randomized study comparing retroperitoneal drainage with no drainage

[27]

[28]

[29] [30]

after lymphadenectomy in gynecologic malignancies. Gynecol Oncol 1997;65: 478–82. Kim HY, Kim JW, Kim SH, Kim YT, Kim JH. An analysis of the risk factors and management of lymphocele after pelvic lymphadenectomy in patients with gynecologic malignancies. Cancer Res Treat 2004;36:377–83. Achouri A, Huchon C, Bats AS, Bensaid C, Nos C, Lecuru F. Complications of lymphadenectomy for gynecologic cancer. Eur J Surg Oncol 2013;39:81–6. Conte M, Panici PB, Guariglia L, Scambia G, Greggi S, Mancuso S. Pelvic lymphocele following radical para-aortic and pelvic lymphadenectomy for cervical carcinoma: incidence rate and percutaneous management. Obstet Gynecol 1990;76:268–71. Petru E, Tamussino K, Lahousen M, Winter R, Pickel H, Haas J. Pelvic and paraaortic lymphocysts after radical surgery because of cervical and ovarian cancer. Am J Obstet Gynecol 1989;161:937–41. Benedet JL, Turko M, Boyes DA, Nickerson KG, Bienkowska BT. Radical hysterectomy in the treatment of cervical cancer. Am J Obstet Gynecol 1980;137:254–62. Yamamoto R, Saitoh T, Kusaka T, Todo Y, Takeda M, Okamoto K, et al. Prevention of lymphocyst formation following systematic lymphadenectomy. Jpn J Clin Oncol 2000;30:397–400. Cibula D, Abu-Rustum NR. Pelvic lymphadenectomy in cervical cancer–surgical anatomy and proposal for a new classification system. Gynecol Oncol 2010;116: 33–7. Fischerova D. Ultrasound scanning of the pelvis and abdomen for staging of gynecological tumors: a review. Ultrasound Obstet Gynecol 2011;38:246–66. Timmerman D, Valentin L, Bourne TH, Collins WP, Verrelst H, Vergote I, et al. Terms, definitions and measurements to describe the sonographic features of adnexal tumors: a consensus opinion from the International Ovarian Tumor Analysis (IOTA) Group. Ultrasound Obstet Gynecol 2000;16:500–5. Ilancheran A, Monaghan JM. Pelvic lymphocyst—a 10-year experience. Gynecol Oncol 1988;29:333–6. Logmans A, Kruyt RH, de Bruin HG, Cox PH, Pillay M, Trimbos JB. Lymphedema and lymphocysts following lymphadenectomy may be prevented by omentoplasty: A pilot study. Gynecol Oncol 1999;75:323–7. Soliman PT, Frumovitz M, Spannuth W, Greer MJ, Sharma S, Schmeler KM, et al. Lymphadenectomy during endometrial cancer staging: practice patterns among gynecologic oncologists. Gynecol Oncol 2010;119:291–4. Tam KF, Lam KW, Chan KK, Ngan HY. Natural history of pelvic lymphocysts as observed by ultrasonography after bilateral pelvic lymphadenectomy. Ultrasound Obstet Gynecol 2008;32:87–90. Hedgire SS, Pargaonkar VK, Elmi A, Harisinghani AM, Harisinghani MG. Pelvic nodal imaging. Radiol Clin North Am 2012;50:1111–25. McMahon CJ, Rofsky NM, Pedrosa I. Lymphatic metastases from pelvic tumors: anatomic classification, characterization, and staging. Radiology 2010;254: 31–46. Lengele B, Scalliet P. Anatomical bases for the radiological delineation of lymph node areas. Part III: Pelvis and lower limbs. Radiother Oncol 2009;92:22–33. Ghezzi F, Uccella S, Cromi A, Bogani G, Robba C, Serati M, et al. Lymphoceles, lymphorrhea, and lymphedema after laparoscopic and open endometrial cancer staging. Ann Surg Oncol 2012;19:259–67. Franchi M, Trimbos JB, Zanaboni F v d, Velden J, Reed N, Coens C, et al. Randomised trial of drains versus no drains following radical hysterectomy and pelvic lymph node dissection: a European Organisation for Research and Treatment of CancerGynaecological Cancer Group (EORTC-GCG) study in 234 patients. Eur J Cancer 2007;43:1265–8. Gallotta V, Fanfani F, Rossitto C, Vizzielli G, Testa A, Scambia G, et al. A randomized study comparing the use of the Ligaclip with bipolar energy to prevent lymphocele during laparoscopic pelvic lymphadenectomy for gynecologic cancer. Am J Obstet Gynecol 2010;203(483):e1–6. Rohaizak M, Khan FJ, Jasmin JS, Mohd Latar NH, Abdullah SS. Ultracision versus electrocautery in performing modified radical mastectomy and axillary lymph node dissection for breast cancer: a prospective randomized control trial. Med J Malaysia 2013;68:204–7. Abu-Rustum NR. Sentinel lymph node mapping for endometrial cancer: a modern approach to surgical staging. J Natl Compr Canc Netw 2014;12:288–97. Cibula D, Abu-Rustum NR, Dusek L, Slama J, Zikan M, Zaal A, et al. Bilateral ultrastaging of sentinel lymph node in cervical cancer: Lowering the false-negative rate and improving the detection of micrometastasis. Gynecol Oncol 2012;127: 462–6.

A prospective study examining the incidence of asymptomatic and symptomatic lymphoceles following lymphadenectomy in patients with gynecological cancer.

To identify the incidence of asymptomatic and symptomatic (i.e., causing pain, hydronephrosis, venous thrombosis, acute lymphedema of the lower or uri...
331KB Sizes 0 Downloads 11 Views