Prognostic value of lymphovascular space invasion determined with hematoxylin–eosin staining in early stage cervical carcinoma: results of a multivariate analysis

P. Morice1,+, P. Piovesan1, A. Rey2, D. Atallah1, C. Haie-Meder3, P. Pautier4, L. Sideris1, C. Pomel1, P. Duvillard5 and D. Castaigne1

Departments of 1 Surgery, 2 Biostatistics, 3 Radiation Therapy and 4 Medical Oncology, 5 Pathology, Institut Gustave Roussy, Villejuif, France

Received 6 February 2003; revised 9 May 2003; accepted 17 June 2003


    Abstract
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Background:

Our aim was to study the prognostic value of the presence of lymphovascular space invasion (LVSI) in patients with stage IB and II cervical carcinoma treated by initial surgery.

Patients and methods:

A retrospective analysis was performed on 193 patients who underwent, between 1985 and 1998, an initial radical hysterectomy with pelvic (± para-aortic) lymphadenectomy using midline laparotomy for stage IB (180 patients) or II (13 patients) cervical carcinoma. Postoperative therapy was delivered according to prognostic factors.

Results:

The rate of LVSI correlated significantly with tumor stage, nodal status and the location of positive nodes. Using univariate analysis, tumor size (<4 or >=4 cm), LVS status and nodal status were prognostic factors. At multivariate analysis, two prognostic factors were identified: LVS status and nodal status. In a subgroup of 89 patients with a small tumor (<=2 cm) and absence of nodal or isthmic involvement, the overall survival was significantly correlated with the presence of LVSI.

Conclusions:

LVSI is a frequent occurrence in patients with early stage cervical cancer. It represents an unfavorable prognostic factor in univariate and multivariate analyses. Such results suggest that improvement is needed in the treatment of patients with a small tumor and LVS invasion.

Key words: cervical cancer, multivariate analysis, nodal involvement, vascular space invasion


    Introduction
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Several prognostic factors have been clearly identified in patients with cervical carcinoma: stage of the disease, tumor size and nodal status. The presence of one of the above prognostic factors highlights the need to improve treatment modalities. Other factors remain disputable and thus have no influence on treatment: histological subtypes (adenocarcinoma versus squamous cell carcinoma), tumor differentiation and patient’s age. However, the presence of lymphovascular space invasion (LVSI) had a prognostic impact, as debated in the literature. Several studies have demonstrated that the presence of LVSI is an independent prognostic factor [1, 2]. Nevertheless, other series failed to demonstrate such results [35]. Recently, Creasman reported an interesting literature review on 18 papers, which included a multivariate analysis about this issue [6]. He concluded that LVSI does not appear to be an independent prognostic factor and therefore should not be a consideration with regard to determination of optimal therapy .

In order to evaluate the prognostic impact and therapeutic implications of the LVS status, we decided to conduct a retrospective analysis of a large number of patients treated in our institution by initial radical surgery since 1985 for stage IB and II cervical carcinoma.


    Patients and methods
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
From May 1985 to May 1998, 577 patients were treated by surgery combined with radiotherapy for invasive stage IB or II cervical carcinoma according to the 1995 International Federation of Gynecology and Obstetrics classification. All patients underwent type III radical hysterectomy plus pelvic ± para-aortic lymphadenectomy by midline laparotomy. Tumor size was determined during the clinical examination. Para-aortic lymphadenectomy was performed in all patients from 1985 to 1994. Since 1995 this procedure was performed in patients with tumor size >3 cm and in patients with tumor size <=3 cm but with positive pelvic nodes during frozen section analysis [7, 8]. In our institution, patients with cervical cancer were usually treated during this period with pre-operative brachytherapy (tumor <5 cm) or with pre-operative external radiation therapy (stage II and/or tumor size >=5 cm). Initial surgery was proposed in two selected cases: first, mainly in young patients (aged <40 years) with early stage disease in order to preserve ovarian function using an ovarian preservation and transposition or, secondly, in patients with suspicious adnexal disease associated with their cervical tumor. During the study period, 193 patients underwent initial surgery. These patients comprised our study group. Prognostic factors and LVS status were studied in this subgroup of patients. All pathological examinations were reviewed by the same pathologist. LVS status was determined using hematoxylin–eosin (H&E)-stained sections. LVS was defined by the presence of tumor cells in the luminal space lined by endothelial cells (Figure 1).



View larger version (149K):
[in this window]
[in a new window]
 
Figure 1. Histological aspect of lymphovascular space invasion (hematoxylin eosin stained; x100).

 
Postoperative treatment depended on tumor stage and size, LVS status, nodal status and the number and topography of nodal involvement. In patients with a very small tumor (<1 cm), absence of parametrial or nodal involvement and free margins, no postoperative treatment was delivered. In patients with a tumor size <4 cm (with or without LVS in the cervix) and absence of nodal or parametrial involvement, a postoperative vaginal brachytherapy [60 Gy according to the International Commission on Radiation Units and Measurements (ICRU) recommendations] [9] was performed.

Postoperative pelvic irradiation (40–45 Gy), followed by vaginal brachytherapy (15–20 Gy according to the ICRU recommendations) [9], was delivered to patients with a bulky tumor (extension to the uterus and/or involvement of parametria) and/or to patients with histologically proven pelvic node involvement and/or to patients with LVSI in parametria or numerous LVSI in the paracervix or paravagina. Since 1996, external radiation therapy has been delivered with concomitant platinum-based chemotherapy (40 mg/m2/weekly) to patients with nodal or parametrial involvement. Pelvic irradiation (40–45 Gy) and cisplatin-containing chemotherapy (100 mg/m2/4 weeks) were given to patients with metastatic common iliac and/or para-aortic nodes. Postoperative para-aortic irradiation was not performed after complete para-aortic lymphadenectomy.

Statistical analysis
The {chi}2 test was used to compare percentages and P <0.05 was considered statistically significant. Survival curves were calculated using the Kaplan–Meier method. Groups were compared in a univariate analysis using the log-rank test and a multivariate analysis was conducted using Cox’s model.


    Results
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
The median age of patients was 37 years (range 21–74). Stages of disease were IB in 180 patients and II in 13 patients (IIA in six and IIB in seven patients). Tumor size was <=2 cm in 126 (65%) patients, between 2 and 4 cm in 40 (21%) patients and >4 cm in 27 (14%) patients. Disease spread, determined during the histological examination, was located in the endocervix in 67 (35%) patients, in the uterine isthmus in 49 (25%) patients and in the uterine corpus in 23 (12%) patients. Nodal involvement in the pelvic nodes was discovered in 48 (25%) patients (with positive para-aortic nodes in eight). LVSI was discovered in 87 (45%) patients. Patient characteristics according to LVS status are detailed in Table 1.


View this table:
[in this window]
[in a new window]
 
Table 1. Patients’ characteristics according to the lymphovascular space invasion status
 
Treatments delivered following radical surgery are detailed in Table 2. Patients were followed up for a median duration of 48 (range 1–154) months. Of the 193 patients, 178 are alive and 172 are alive and free of recurrence. Twenty-one patients developed recurrent disease. Sixteen of the 17 patients who died from their tumor had LVSI. Eleven of these recurrent patients had distant metastatic disease (10 of them had LVSI). Two patients (with LVSI) had recurrent disease characterized by metastasis on transposed ovaries.


View this table:
[in this window]
[in a new window]
 
Table 2. Postoperative treatment according to the lymphovascular space status
 
We also studied the outcome in a subgroup of 89 patients with, a priori, a good prognosis (tumor size <=2 cm, free margins with an absence of extension to the uterine isthmus and absence of nodal involvement). In this subgroup, four patients had recurrence of their cervical cancer (three had LVSI). One patient with LVSI was treated for breast cancer associated with cervical tumor that recurred with bone metastases from her breast tumor. This patient was excluded from the following analysis. Location of the recurrent disease was in the pelvic cavity (with para-aortic nodes in one) for two patients, in the vaginal cuff for one patient and in the inguinal nodes for one patient. When we studied the impact of postoperative treatment on the outcomes of 34 patients with LVSI, five patients were treated with exclusive surgery (two of them recurred), 26 received postoperative vaginal brachytherapy (one of them recurred) and three patients received external radiation therapy ± brachytherapy (no recurrence). The only patient with recurrent disease and absence of LVSI was treated with postoperative vaginal brachytherapy. This patient is the only recurrent patient actually alive and disease free.

In the univariate analysis, tumor size >4 cm, LVSI and nodal involvement decreased survival rate significantly (Table 3). The 5-year overall survival was 97% in patients without LVSI compared with 78% for patients with LVSI (P <0.0001) (Figure 2). Disease-free survival at 5 years was 96% in patients without LVSI compared with 74% for patients with LVSI (P <0.0001) (Figure 3). In the subgroup of 145 patients without nodal involvement, overall survival was significantly correlated to LVS status (Figure 4). When we studied the prognostic influence of LVS status in the subgroup of 89 patients with, a priori, a good prognosis (tumor size <=2 cm, free margins with an absence of extension to the uterine isthmus and absence of nodal involvement), the 5-year overall survival was 100% in 55 patients without LVSI compared with 92% in 34 patients with LVSI (P <0.05) (Figure 5). In the multivariate analysis, only two factors were prognostic: LVS status [relative risk (RR) of death 5.7] and nodal involvement (RR of death 4.9 for pelvic positive nodes and 17.1 for para-aortic positive nodes) (Table 4).


View this table:
[in this window]
[in a new window]
 
Table 3. Prognostic factors determined in univariate analysis
 


View larger version (18K):
[in this window]
[in a new window]
 
Figure 2. Overall survival according to the lymphovascular space status. LVS, lymphovascular space.

 


View larger version (18K):
[in this window]
[in a new window]
 
Figure 3. Event-free survival according to the lymphovascular space status. LVS, lymphovascular space.

 


View larger version (18K):
[in this window]
[in a new window]
 
Figure 4. Overall survival according to the lymphovascular space status in 145 patients with negative nodes. LVS, lymphovascular space.

 


View larger version (17K):
[in this window]
[in a new window]
 
Figure 5. Overall survival according to the lymphovascular space status in 87 patients with tumor size £2 cm, absence of involvement of uterine isthmus and negative nodes. LVS, lymphovascular space.

 

View this table:
[in this window]
[in a new window]
 
Table 4. Prognostic factors determined in multivariate analysis
 

    Discussion
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
In this series, LVS status is correlated with nodal status. This status is not correlated with patients’ age, histological subtype or tumor grade. These factors were not clearly identified as independent prognostic factors in cervical cancer. Surprisingly, we did not find any significant correlation between tumor size or stage and LVS status. In our institution, the standard treatment of stage I and II cervical cancer is based on radio-surgical combination. Treatment of patients with tumor size <=2 cm is usually based on preoperative utero-vaginal brachytherapy, followed by radical hysterectomy plus pelvic lymphadenectomy [8, 10]. In some cases, we decided to perform initial surgery before postoperative radiation therapy. The main indication of this management is young patients (aged <40 years) with early stage disease in order to preserve and transpose the ovary before potential brachy-therapy [11]. Since 1988 (when we observed our first case of ovarian metastasis), we have selected this procedure for young patients with a small tumor (<3 cm) and/or stage IB1 disease [12]. This explains why the majority of patients in the present series had a small tumor size and also why the median age of patients was low (37 years). This observation could perhaps be a limitation of the results of the current study. Some patients with stage IB2/II and/or tumor size >3 cm were included in the present series, but most of these patients were treated before 1988. The rates of LVSI seem to be related to tumor size (41% in tumors <2 cm; 48% in tumors between 2 and 4 cm and 59% in tumors >=4 cm), but this difference is not statistically significant, probably because of the low number of patients with tumors >=4 cm.

The present series confirms that the presence of LVSI is an independent prognostic factor in cervical cancer. A Gynecologic Oncology Group (GOG) study, conducted 15 years ago on the prognostic factors for stage IB cervical carcinoma treated by primary radical hysterectomy, concluded that clinical size, depth of tumor invasion and capillary lymphatic space status were independent prognostic factors [1]. The GOG then developed a scoring system including these prognostic factors in order to evaluate the risk of recurrent disease and to define a population who needed adjuvant therapy (particularly in patients with negative nodes) [1]. In the recent review by Creasman involving 18 publications (4000 patients with stage IB or II carcinoma) with multivariate analysis, only three papers found similar results [6]. The therapeutic implications of these results are important. In teams where patients are treated exclusively by radiation therapy for early stage cervical tumor, the presence of LVSI diagnosed during histological examination of (cone) biopsies does not change the modalities of radiation therapy (in the absence of nodal spread). However, for teams where patients are treated by exclusive surgery or radio-surgical combination, our results are important. Since the prognosis of patients with LVSI is worse than the prognosis of patients without LVSI, this treatment should be optimized, and exclusive surgery is not sufficient.

Which treatment optimization, chemotherapy or radiation therapy, should be selected? A recent randomized GOG trial by Sedlis et al. demonstrated that adjuvant external pelvic radiation therapy following radical hysterectomy reduces the number of recurrences, with an acceptable morbidity, in patients with stage IB cervical carcinoma (and negative lymph nodes) associated with stromal invasion of the cervix >1/3 and/or LVSI and/or bulky tumor [13]. Should external radiation be performed in all patients with LVSI? Probably not. External radiation (with concomitant chemotherapy) should be performed in patients with negative nodes but LVSI in the parametria. However, external radiation used as adjuvant therapy following radical hysterectomy increases the rate of complications, particularly for chronic enteritis [14, 15]. Such management (initial radical surgery followed by external radiation therapy) should be avoided in patients with LVS located only in the cervix. In the study by Sedlis et al. brachytherapy was not used. For 30 years such treatment was routinely used in our institution for early stages of cervical carcinoma (combined with radical surgery) and nearly 1000 of these procedures were performed. We observed good survival results using this combination (94% 5-year survival) [8, 10]. The advantage of brachytherapy is that it optimizes local control. No randomized study has demonstrated the superiority of combined brachytherapy plus radical surgery compared with exclusive radical surgery in patients with stage IB1 cervical cancer and negative nodes. However, the good survival results of this radio-surgical combination observed in several retrospective analyses confirm the potential advantage of treatments combining surgery and brachytherapy [10, 14, 16]. In the present series two out of five patients treated by surgery alone with LVSI recurred compared with one out of 26 patients who were treated by radical surgery plus brachytherapy.

Other histological factors are important to evaluate the prognosis of patients with LVS invasion (number and topography of LVSI, type of LVSI: lymphatic or blood vessel invasion) [17, 18]. Such a histological differentiation is difficult and reproducibility in the interpretation of these factors between several pathologists is not yet demonstrated. In order to improve the rate of LVSI detection, should systematic immunostaining be used or is H&E staining sufficient? A recent paper by Obermair et al. demonstrates that the addition of immunostaining analysis for factor VIII-related antigen in order to detect LVSI provides additional information on the outcome of stage IB cervical cancer with negative nodes [19]. However, the cost-effectiveness of this routine procedure compared with H&E staining was not evaluated. Further studies are needed in order to evaluate such results. In the present series, the diagnosis of LVSI was performed by an experienced pathologist using (only) H&E-stained specimens. We observed that none of the patients with small tumors, negative nodes and no LVSI died from recurrent disease. Our criteria to evaluate LVS status using H&E staining are good. The pathologist’s experience is the key point in evaluating this status.

In conclusion, the presence of LVSI is a frequent occurrence in patients with early stage cervical cancer. This is an unfavorable prognostic factor in univariate and multivariate analyses in this series. Such results could suggest that improvement in treatment is needed for patients with small tumors and LVSI. However, further studies are needed in order to evaluate optimal modalities for treatment of patients with early stage tumor associated with LVSI.


    Footnotes
 
+ Correspondence to: Dr P. Morice, Institut Gustave Roussy, 39 rue Camille Desmoulins, 94805 Villejuif, France. Tel: +33-1-42-11-44-39; Fax: +33-1-42-11-52-13; E-mail: morice{at}igr.fr Back


    References
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
1. Delgado G, Bundy B, Zaino R et al. Prospective surgical-pathological study of disease-free interval in patients with stage IB squamous cell carcinoma of the cervix: a Gynecologic Oncology Group study. Gynecol Oncol 1990; 38: 352–357.[ISI][Medline]

2. Sevin BU, Lu Y, Bloch DA et al. Surgically defined prognostic parameters in patients with early cervical carcinoma. A multivariate survival tree analysis. Cancer 1996; 78: 1438–1446.[CrossRef][ISI][Medline]

3. Gauthier P, Gore I, Shingleton HM et al. Identification of histopathologic risk groups in stage IB squamous cell carcinoma of the cervix. Obstet Gynecol 1985; 66: 569–574.[Abstract]

4. Smiley LM, Burke TW, Silva EG et al. Prognostic factors in stage IB squamous cervical cancer patients with low risk for recurrence. Obstet Gynecol 1991; 77: 271–275.[Abstract]

5. McLellan R, Dillon MB, Woodruff JD et al. Long-term follow-up of stage I cervical adenocarcinoma treated by radical surgery. Gynecol Oncol 1994; 52: 253–259.[CrossRef][ISI][Medline]

6. Creasman W. Is lymphovascular space involvement a prognostic factor in early stage cervical cancer? Int J Gynecol Cancer 2002; 12: 602.

7. Michel G, Morice P, Castaigne D et al. Lymphatic spread of stage IB/II cervical carcinoma: anatomy and surgical implications. Obstet Gynecol 1998; 91: 360–363.[Abstract/Free Full Text]

8. Morice P, Castaigne D, Pautier P et al. Interest of para-aortic lymphadenectomy in patients with stage IB and II cervical carcinoma. Gynecol Oncol 1999; 73: 106–110.[CrossRef][ISI][Medline]

9. ICRU Report 38. Dose and volume specification for reporting intracavitary therapy in gynecology. Bethesda, MD: International Commission on Radiation Units and Measurements, 1985.

10. Gerbaulet AL, Kunkler IH, Kerr GR et al. Combined radiotherapy and surgery: local control and complications in early carcinoma of the uterine cervix: the Villejuif experience, 1975–1984. Radiother Oncol 1992; 23: 66–73.[ISI][Medline]

11. Morice P, Juncker L, Rey A et al. Ovarian transposition for patients with cervical carcinoma treated by radio-surgical combination. Fertil Steril 2000; 74: 743–748.[CrossRef][ISI][Medline]

12. Morice P, Haie-Meder C, Pautier P et al. Ovarian metastasis on transposed ovary in patients treated for squamous cell carcinoma of the uterine cervix: report of two cases and surgical implications. Gynecol Oncol 2001; 83: 605–607.[CrossRef][ISI][Medline]

13. Sedlis A, Bundy BN, Rotman MZ et al. A randomized trial of pelvic radiation therapy in selected patients with stage IB carcinoma of the cervix after radical hysterectomy and pelvic lymphadenectomy: a Gynecologic Oncology Group study. Gynecol Oncol 1999; 73: 177–183.[CrossRef][ISI][Medline]

14. Atlan D, Touboul E, Deniaud-Alexandre E et al. Operable stages IB and II cervical carcinomas: a retrospective study comparing preoperative uterovaginal brachytherapy and postoperative radiotherapy. Int J Radiat Oncol Biol Phys 2002; 54: 780–793.[ISI][Medline]

15. Guillermin P, Morice P, Rey A et al. Influence de la séquence thérapeutique dans la survenue des entérites radiques chez les patientes traitées pour un cancer utérin ou vaginal. 13th Congrès de la Société Française d'Oncologie Gynécologique. Paris. 21–23 November 2002.

16. Resbeut MR, Alzieu C, Gonzague-Casabianca L et al. Combined brachytherapy and surgery for early carcinoma of the uterine cervix: analysis of extent of surgery on outcome. Int J Radiat Oncol Biol Phys 2001; 50: 873–881.[ISI][Medline]

17. Sakuragi N, Takeda N, Hareyama H et al. A multivariate analysis of blood vessel and lymph vessel invasion as predictors of ovarian and lymph node metastases in patients with cervical carcinoma. Cancer 2000; 88: 2578–2583.[CrossRef][ISI][Medline]

18. Roman LD, Felix JC, Muderspach LI et al. Influence of quantity of lymph-vascular space invasion on the risk of nodal metastases in women with early-stage squamous cancer of the cervix. Gynecol Oncol 1998; 68: 220–225.[CrossRef][ISI][Medline]

19. Obeirmair A, Wanner C, Bilgi S et al. The influence of vascular space involvement on the prognosis of patients with stage IB cervical carcinoma. Correlation of results from hematoxylin and eosin staining with results from immunostaining for factor VIII-related antigen. Cancer 1998; 82: 689–696.[CrossRef][ISI][Medline]





This Article
Abstract
Full Text (PDF)
E-letters: Submit a response
Alert me when this article is cited
Alert me when E-letters are posted
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (7)
Disclaimer
Request Permissions
Google Scholar
Articles by Morice, P.
Articles by Castaigne, D.
PubMed
PubMed Citation
Articles by Morice, P.
Articles by Castaigne, D.