Abstract
Aim: To assess the feasibility of less radical surgery in patients with incidentally-detected cervical cancer at simple hysterectomy. Patients and Methods: High-risk factors (HFs) were evaluated in 104 patients who underwent radical hysterectomy and lymphadenectomy for the treatment of stage IA1-IB1 cervical cancer according to intermediate-risk factors (IFs). Results: Thirty-three patients without IFs had no HFs. As the number of IFs increased, the number of HFs increased (p=0.009). Multivariate analysis revealed that lymph node metastasis was the only independent risk factor for parametrial involvement (hazard ratio, 31.3; 95% confidence interval, 1.6–599.4; p=0.022). An absence of IFs was associated with a longer progression-free survival than the presence of IFs in the subgroup analysis of favorable histologies (p=0.044). Conclusion: HFs could be excluded in stage IA1-IB1 cervical cancer without IFs. Omitting parametrectomy seems a feasible option for selected patients with incidentally-detected early-stage cervical cancer at simple hysterectomy, without IFs.
Invasive cervical cancer incidentally detected after simple hysterectomy has been shown to have a poor prognosis without adjuvant treatment (1). Therefore, clinical practice guidelines recommend either radiation therapy with concurrent chemotherapy or complete parametrectomy with upper vaginectomy and pelvic lymph node (LN) dissection with para-aortic LN sampling, as an adjuvant therapy. These recommendations are made even if the resection margin is negative and no abnormal finding is observed in imaging studies (2).
However, there is accumulating evidence suggesting that parametrectomy could safely be omitted in patients with small early-stage cervical carcinomas. This is based on the very low rate of parametrial involvement in some patients and may reduce the morbidity of parametrectomy without compromising the oncological outcome (3). Multiple studies have provided criteria useful for identifying patients with a low risk (≤1%) of parametrial involvement. The most consistently reported criteria include tumor size ≤2 cm and a favorable histology, such as squamous cell carcinoma (SCC), adenocarcinoma (AC), and adenosquamous cell carcinoma (AS) (4-6). A stromal invasion depth of ≤5 mm or 10 mm and no lymphovascular space invasion (LVSI) were frequently reported as useful prognostic factors for identifying candidates for less radical surgery (7, 8). A randomized trial comparing radical hysterectomy and pelvic node dissection versus simple hysterectomy and pelvic node dissection in patients with low-risk early-stage cervical cancer (SHAPE), an ongoing Gynecologic Cancer Intergroup clinical trial led by the National Cancer Institute (NCI) of Canada, aims to demonstrate that simple hysterectomy is as effective as radical hysterectomy in selected women with early-stage cervical cancer (9). The inclusion criteria of the SHAPE trial were histologically confirmed early-stage SCC, AC, or AS of the cervix with low-risk features such as a stromal invasion <10 mm and a tumor size ≤2 cm, irrespective of the presence or absence of LVSI.
Of note, the criteria of low-risk features include intermediate-risk factors (IFs), all of which can be measured adequately from simple hysterectomy specimens. Our group recently reported that a low-risk group of patients in whom disease was unlikely to recur without adjuvant treatment could be identified by using the combination of IFs measured from simple hysterectomy specimens in patients with cervical cancer found after such a procedure (10). We therefore considered that less radical surgery could be used in selected cases of incidentally-found cervical cancer after simple hysterectomy. This study aimed to assess the feasibility of less radical surgery in such patients through the analysis of IFs in radical hysterectomy specimens.
Patients and Methods
Patients. We retrospectively reviewed the medical records of 116 patients who underwent primary radical hysterectomy and pelvic LN dissection for the treatment of stage IA1-IB1 cervical cancer, according to the International Federation of Gynecology and Obstetrics (FIGO) (2), between 2003 and 2010. Twelve patients with stage IA1 and no LVSI were excluded as simple hysterectomy alone is a definitive treatment without any safety concern for this group. No patient received neoadjuvant chemotherapy before radical hysterectomy. Clinicopathological characteristics of the remaining 104 patients were included in the analysis (Table I). Their median age was 49.5 years (range=24-77 years). Three patients with FIGO stage IA1 underwent RH based on the finding of LVSI from cervical conization specimens. Ninety-eight (94.2%) had favorable histologies such as SCC, AC, and AS. During the median follow-up period of 53 months (range=0-119 months), there were nine cases of recurrence (8.7%) and five cases of disease-specific mortality (4.8%).
Intermediate- and high-risk prognostic factors. IFs included tumor size >2 cm, stromal invasion >5 mm, and LVSI. Correlations between IFs and high-risk factors (HFs) are shown in Table II. Of note, specificities of 100% for parametrial involvement were remarkable in both measurements of stromal invasion in millimeters and trisection, as well as LVSI. Pathological tumor size was used, rather than clinical or radiological tumor size, as this is an important factor in deciding adjuvant treatment for incidentally detected cervical cancer at simple hysterectomy. The mean tumor size was 2.1±1.1 cm (range=0.1-4.0 cm). Stromal invasion both in millimeters (5 mm) and trisection (>2/3) showed excellent correlations with HFs, except resection margin involvement. Nevertheless, we used the criterion of stromal invasion >5 mm as there were more recurrent cases in trisection measurement compared with millimeter measurement among 78 patients with stromal invasion <2/3 (3.8% versus 1.9%). The mean depth of stromal invasion was 7.14±5.74 mm (range=1.00-22.0 mm). HFs included parametrial involvement, resection margin involvement, and LN metastasis. For the 15 patients with LN metastasis, the mean number of LNs retrieved and LNs positive for malignancy were 26.8±7.1 (range=14-41) and 2.5±2.0 (range=1-7), respectively.
Adjuvant treatment. Of 13 patients who received adjuvant concurrent chemoradiation therapy (CCRT), 9 (69.2%), 3 (23.1%), and 1 (7.7%) received paclitaxel/carboplatin, fluorouracil/cisplatin, and weekly cisplatin during RT, respectively. External irradiation was delivered to the whole pelvis at 1.8 or 2.0 Gy per fraction once daily, five days per week. The median dose to the whole pelvis was 50.4 Gy (range=45.0-50.4 Gy). Seventy-eight patients received no adjuvant treatment after simple hysterectomy.
Statistical analysis. IFs were univariately evaluated as a function of HFs using a chi-square test and Fisher's exact test where appropriate, calculating the hazard ratio (HR) with 95% confidence interval (CI) or p-value. Moreover, IFs and other prognostic factors were multivariately assessed for every HF using a multiple logistic regression analysis. According to the number of IFs, the ratio of the patients with HFs and the mean number of HFs per person was evaluated using a chi-square test and analysis of variance test, respectively. Progression-free survival (PFS) rates were measured from the date of surgery to the date of recurrence. The PFS rates between patients with and without IFs were compared using Kaplan–Meier survival analysis with log-rank test. We performed a subgroup analysis of the 98 patients with favorable histologies. Statistical tests were two-sided, with p<0.05 indicating significance. Statistical analyses were performed using SPSS for Windows, version 19.0 (SPSS Inc., Chicago, IL, USA). The Institutional Review Board of Seoul National University Bundang Hospital approved this study (B-0906/078-102).
Patients' characteristics (n=104).
Correlation between high- and intermediate-risk factors in early-stage cervical cancer.
Results
No IF, no HF. Tumor size >2 cm, stromal invasion >5 mm, and LVSI were found in 53 (51.0%), 51 (49.0%), and 39 (37.5%) patients, respectively. Thirty-three (31.7%) patients had neither an IF nor a HF. As the number of IFs increased, the ratio of patients with HFs significantly increased (Figure 1A). Moreover, a trend for an increase in the mean number of HFs according to the number of IFs was observed (p=0.058) (Figure 1B), again noting that HFs were not present without IFs. However, survival analysis revealed that having no IF was not significantly associated with a longer PFS than having one or more IF (p=0.170) (Figure 2A).
Association of HFs with clinicopathological factors, including IFs. Parametrial involvement was the only independent prognostic factor for poor PFS among the six risk factors (HR=76.5; 95% CI=4.5-1291.8; p=0.003) (data not shown). Univariate analysis revealed that parametrial involvement was significantly associated with all three IFs (Table II), although LN metastasis was the only independent risk factor for parametrial involvement on multivariate analysis (HR=31.3; 95% CI=1.6-599.4; p=0.022) (Table III). However, multivariate analysis revealed that parametrial involvement (HR=23.0 95% CI=1.7–305.6; p=0.017) and LVSI (HR=6.9; 95% CI=1.2-41.6; p=0.034) were independent prognostic factors for LN metastasis (Table IV). No clinicopathological factor evaluated in our study was associated with resection margin involvement in both univariate and multivariate analyses.
A subgroup analysis of patients with favorable histologies. Among 98 patients with favorable histologies, no patient had HFs without IFs. Univariate and multivariate associations between IFs and HFs were similar to those for the whole study population. However, the association of having no IF with a longer PFS reached a statistical significance on survival analysis of the subgroup of patients with favorable histologies (p=0.044) (Figure 2B). More patients had two or more IFs in cases of unfavorable histology compared to those with favorable histology (83.3% versus 43.9%). Furthermore, more patients had one or more HF in cases of unfavorable histology compared to those with favorable histology (33.3% versus 15.3%). Notably, one patient with small cell carcinoma who did not receive adjuvant treatment due to presenting with neither an IF nor a HF eventually showed recurrence during follow-up.
Discussion
We demonstrated, through the present study, that no patient with early-stage cervical cancer had HFs without IFs. As the number of IFs increased, the risk of having an HF significantly increased. The significant difference in PFS between patients with and those without IFs was observed only for those with favorable histologies such as SCC, AC, and AS. LN metastasis was the only independent risk factor for parametrial involvement, which was itself the only independent prognostic factor for poor PFS among the six risk factors.
Adjuvant treatment after radical hysterectomy was decided based on the presence of risk factors. If HFs are present, the recurrence rate increases to 40% (11, 12). Based on the promising survival outcomes after CCRT in clinical trials, CCRT with weekly intravenous cisplatin was established as a new standard treatment for high-risk early-stage cervical cancer (13-15). However, the clinical significance of IFs in managing early-stage cervical cancer is still controversial. While the current standard regimen for patients with two or more IFs after radical hysterectomy is adjuvant RT (16, 17), some physicians believe RT is sufficient considering the relatively favorable prognosis of patients with only IFs (85-90% 5-year survival) who do not receive any other adjuvant treatment (18). However, a Gynecologic Oncology Group (GOG) randomized phase III clinical trial of adjuvant RT versus CCRT in patients with ‘high intermediate-risk’ stage I-IIA cervical cancer is currently underway (GOG-263) (Figure 3), based on the considerable risk of recurrence (2-31%) at three years in this group of patients (17, 19, 20). Further, large prospective clinical trials of less radical surgery in patients with ‘very low intermediate-risk’ stage I-IB1 cervical cancer are ongoing worldwide mainly by GOG and the NCI of Canada. The concept of less radical surgery in patients with ‘very low intermediate-risk’ cancer might be reasonably applied to the management of cervical cancer incidentally detected after simple hysterectomy (Figure 3), as the purpose of these trials is to demonstrate the safety of simple hysterectomy in this group of patients compared with radical hysterectomy (9).
Univariate and multivariate analyses of clinicopathological risk factors as a function of parametrial involvement in patients with early-stage cervical cancer.
Univariate and multivariate analyses of clinicopathological risk factors as a function of lymph node metastasis in patients with early-stage cervical cancer.
Invasive cervical cancer incidentally found after simple hysterectomy is not common enough for a large-scale clinical study, and no definitive data are available to guide the appropriate adjuvant treatment of these patients (2). This urged us try to assess the risk of recurrence in this group of patients using a larger pool of patients undergoing radical hysterectomy. HFs, such as parametrial involvement, can be evaluated precisely in specimens from radical hysterectomy with adequate parametrial tissue, but not in specimens from simple hysterectomy (10). However, all IFs are available for evaluation in specimens from simple hysterectomy. Therefore, if the risk of a HF from a specimen from radical hysterectomy was predicted by an IF in a specimen from simple hysterectomy, patients who underwent radical hysterectomy for the treatment of early-stage cervical cancer could become an appropriate study population for defining patients who may need adjuvant treatment after simple hysterectomy.
Correlation between number of intermediate-risk and high-risk factors. A significant increase of the ratio of patients with high-risk factors (p=0.009) (A) and the mean number of high-risk factors per person (p=0.058) (B) according to the number of intermediate-risk factors.
Comparison of progression-free survival of patients with early-stage cervical cancer between those with and those without intermediate-risk factors using Kaplan–Meier survival analysis with log-rank test in the overall study population (A) and patients with favorable histology (B).
The associations between IFs and HFs have been previously reported, although they seem somewhat inconsistent. Clinical tumor size >4 cm (19), deep stromal invasion (20), and LVSI (21, 22) were shown to be associated with a significant increase in LN metastasis, as well as poor survival outcomes. Boyce et al. also reported that LVSI contributed to prognostic information beyond that available from lesion size (23). However, Roche and Norris questioned the validity of the pathological finding of lymphatic invasion, because the pattern of invasion within the stroma did not correlate with the presence or absence of lymphatic invasion, noting that all of the LNs with this pattern of invasion in microinvasive SCC were negative for metastasis (24).
To the best of our knowledge, this is the first study defining a low-risk group for less radical surgery, omitting parametrectomy and upper vaginectomy, in incidentally-found early-stage cervical cancer after simple hysterectomy, using IFs in radical hysterectomy pathology. No patient had HFs without IFs. It is noteworthy that selection criteria used to define the eligible population for non-radical surgery in several study groups often included negativity for LN metastasis (3). Kim et al. reported 3.6% of patients with LN metastasis even when their criterion of stromal invasion ≤5 mm was met (7). Therefore, our suggestion for adjuvant treatment for small-sized early-stage cervical cancer incidentally found after simple hysterectomy, with both negative resection margin and IFs, would be pelvic LN dissection with or without para-aortic LN sampling alone. Despite the originality of the study concept, the relatively small sample size and retrospective design were limitations. In addition, the possibility that the conclusion may differ for a different population is another limitation. Most relevant studies adopted clinical tumor size; however, we used the pathological tumor size for our criteria because a clinical tumor size before surgery was not available for incidentally-found cervical cancer.
The application of the concept of less-radical-surgery to incidentally-found cervical cancer after simple hysterectomy (SH) and its rationale in viewing of the management of early-stage cervical cancer treated by radical hysterectomy (RH) according to intermediate-risk factors (IF). CCRT, Concurrent chemoradiation therapy; DOI, depth of invasion; GOG, Gynecologic Oncology Group; LVSI, lymphovascular space invasion; PLND, pelvic lymph node dissection; RT, radiation therapy.
Briefly, in patients with early-stage cervical cancer without IFs, who underwent RH and pelvic LN dissection, no HF was found. The significant difference in PFS between patients with and those without IFs was observed only for patients with favorable histologies such as SCC, AC, and AS. In conclusion, pelvic LN dissection with or without para-aortic LN sampling alone could be a feasible option for patients with stage 1A-1B1 invasive cervical cancer incidentally detected after simple hysterectomy, if pathology reports negative resection margin, favorable histology, and absence of all three intermediate-risk factors: tumor size ≥2 cm, LVSI, and stromal invasion ≥5 mm. A prospective study with a corresponding study population is warranted.
- Received September 16, 2013.
- Revision received October 18, 2013.
- Accepted October 22, 2013.
- Copyright© 2013 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.