Prognostic factors in neuroendocrine small cell cervical carcinoma

A multivariate analysis

Authors

  • John K. Chan M.D.,

    1. Division of Gynecologic Oncology, The Chao Family Comprehensive Cancer Center, University of California–Irvine Medical Center, Orange, California
    Search for more papers by this author
  • Vera Loizzi M.D.,

    1. Division of Gynecologic Oncology, The Chao Family Comprehensive Cancer Center, University of California–Irvine Medical Center, Orange, California
    Search for more papers by this author
  • Robert A. Burger M.D.,

    1. Division of Gynecologic Oncology, The Chao Family Comprehensive Cancer Center, University of California–Irvine Medical Center, Orange, California
    Search for more papers by this author
  • Joanne Rutgers M.D.,

    1. Department of Pathology, Long Beach Memorial Medical Center, Long Beach, California
    Search for more papers by this author
  • Bradley J. Monk M.D.

    Corresponding author
    1. Division of Gynecologic Oncology, The Chao Family Comprehensive Cancer Center, University of California–Irvine Medical Center, Orange, California
    • Division of Gynecologic Oncology, The Chao Family Comprehensive Cancer Center, University of California–Irvine Medical Center, 101 The City Drive, Orange, CA 92868
    Search for more papers by this author
    • Fax: (714) 456-6463


Abstract

BACKGROUND

The purpose of this study was to evaluate the clinical and pathologic factors associated with survival in patients with neuroendocrine (NE) cervical carcinoma.

METHODS

All patients with NE cervical carcinoma diagnosed between 1979–2001 were identified from tumor registry databases at two hospitals. Data were collected from hospital charts, office records, and tumor registry files. The impact of clinical and pathologic risk factors on the survival of patients with small cell NE carcinoma of the cervix was evaluated using Kaplan–Meier life table analyses and log-rank tests. The independent prognostic factors found to be predictive of survival in univariate analysis were evaluated using Cox regression. All tests were two-tailed with P values < 0.05 considered significant.

RESULTS

Thirty-four patients (median age, 42 years) were diagnosed with neuroendocrine cervical carcinoma, which included 21 with International Federation of Gynecology and Obstetrics (FIGO) Stage I disease, 6 with FIGO Stage II disease, 5 with FIGO Stage III disease, and 2 with FIGO Stage IV disease. Seventeen patients underwent a radical and 6 patients underwent a simple hysterectomy. Fourteen women received adjuvant therapy with pelvic radiation and/or cisplatin-based chemotherapy. Ten women received primary radiotherapy with (n = 5) or without (n = 4) chemotherapy and the remaining patient refused therapy. Women with early-stage (Stage I-IIA) disease had median survival rates of 31 months compared with 10 months in the advanced-stage (Stage IIB-IVB) group (P = 0.002). In univariate analysis, advanced stage (P = 0.002), tumor size >2 cm (P = 0.02), margin involvement (P = 0.016), pure versus a mixed histologic pattern (P = 0.04), margin status (P = 0.016), and smoking (P = 0.04) were considered poor prognostic factors. In multivariate analysis, smoking for early-stage patients and stage of disease in the overall population remained as independent prognostic factors of survival.

CONCLUSIONS

Smoking and advanced stage are reported to be poor prognostic factors for survival in patients with NE small cell carcinoma of the cervix. Only those with early lesions amenable to extirpation are cured. The role of primary or postoperative radiation with or without chemotherapy is unclear and yields uniformly poor results, particularly in patients with advanced lesions. Clinical trials are needed. Cancer 2003;97:568–74. © 2003 American Cancer Society.

DOI 10.1002/cncr.11086

Small cell neuroendocrine (NE) carcinoma is a rare tumor that comprises 1–6% of cervical tumors.1, 2 Under light and electron microscopy, the characteristics of the tumor are indistinguishable from oat cell carcinoma of the lung. Previous reports have shown that women diagnosed with small cell carcinoma of the cervix have greater frequency of lymph node metastases, lymphovascular invasion, recurrence, and poorer prognosis compared with those with other types of cervical malignancies.1, 2 As one of the more aggressive gynecologic malignancies, the 5-year survival rates for patients with this tumor has been reported as 14% in advanced stages.3 Thus, small cell NE cervical carcinoma provides a therapeutic challenge for the oncologist. Given the poor prognosis, it is important to identify prognostic factors responsible for survival in an effort to improve treatment strategies. In fact, the Gynecologic Oncology Group (GOG) attempted to study NE carcinoma of the cervix in Protocol 66 between 1982–1986 but failed to accrue enough patients for this trial. In this current series, patients diagnosed with small cell NE carcinoma of the cervix were identified and the clinical and pathologic prognostic factors responsible for survival were determined. Furthermore, various therapeutic strategies in the treatment of this aggressive tumor were investigated.

MATERIALS AND METHODS

All patients with NE cervical carcinoma diagnosed between 1979–2001 were identified from tumor registry databases at University of California–Irvine Medical Center and Long Beach Memorial Medical Center. Data were collected from hospital charts, office records, and tumor registry files. All patients underwent a physical examination, chest roentgenogram, intravenous pyelography, cystoscopy, and proctoscopy when indicated. These women then were staged based on the International Federation of Gynecology and Obstetrics (FIGO) clinical staging criteria for cervical carcinoma regardless of their operative findings. Early stage was defined as Stages I-IIA and advanced stage as Stages IIB-IVB, with the majority of patients diagnosed with early-stage disease. Patients with tumors confined to the cervix typically underwent a radical hysterectomy with pelvic lymphadenectomy; whereas women with advanced tumors were treated with radiotherapy with or without chemotherapy.

All pathologic specimens were reviewed at multidisciplinary tumor board involving gynecologic oncologists, pathologists, and radiation oncologists from both institutions. The hysterectomy specimens were examined macroscopically to determine the size of the tumors. Small tumors were defined as those <2 cm in its greatest dimension and large tumors as those >2 cm. By light microscopy, all tumors were diagnosed as small cell NE cervical carcinoma. Immunohistochemical stains using standard techniques were performed in 10 cases and electron microscopy in 24 cases to confirm the diagnosis. These tumor cells appeared small, round, and oat-shaped with a high nuclear to cytoplasmic ratio. In addition, the nucleolus is not easily noted and the nucleus often is described as having a “salt and pepper” appearance.4

Statistical Analysis

Statistical analysis was performed using NCSS 2001.5 The impact of clinical and pathologic risk factors on the survival of patients with small cell NE carcinoma of the cervix was evaluated using Kaplan–Meier life table analyses and log-rank tests. The independent prognostic factors found to be predictive of survival in univariate analyses were evaluated using Cox regression. All tests were two-tailed with P values < 0.05 considered significant.

RESULTS

Of the 34 patients, 19 had Stage IB1, 2 had Stage IB2, 4 had Stage IIA, 2 had Stage IIB, 5 had Stage III, and 2 had Stage IV disease. Demographic characteristics of the patients are shown in Table 1. The most common presenting symptom was vaginal bleeding. Among the 23 women treated with hysterectomy, 17 underwent a radical and 6 underwent a simple hysterectomy. Of the six women who underwent a simple hysterectomy, three were found to have extensive lymph node invasion during surgery. The remaining three patients were referred to the University Hospital after undergoing surgery at an outside institution and were found to have an unsuspected tumor of the cervix. Approximately 68% of the patients (23 of 34 patients) underwent a pelvic and/or paraaortic lymphadenectomy. Three patients with locally advanced disease underwent lymph node sampling only to determine the extent of disease. Given the aggressive behavior of this tumor, an increasing number of women received adjuvant therapy during the latter part of the study. In fact, 15 women received postoperative therapy including 11 with radiation, 3 with chemoradiation, and 1 with chemotherapy only.

Table 1. Demographic Characteristics (n = 34)
VariablesNo. or %
  1. HRT: hormone replacement therapy.

Age (yrs) (median) (range)42 (28–79)
Gravity (median) (range)2 (0–9)
Parity (median) (range)2 (0–8)
Race 
 White68% (23/34)
 Black6% (2/34)
 Hispanic6% (2/34)
 Others21% (7/34)
Menopausal35% (12/34)
HRT21% (7/34)
Smoking24% (8/34)

Moreover, patients with advanced disease typically underwent radiotherapy with or without chemotherapy. The 11 women who did not undergo a hysterectomy included 1 with Stage IB2 disease, 3 with Stage II disease, 5 with Stage III disease, and 2 with Stage IV disease. Of these patients, five underwent cisplatin-based chemoradiation, four underwent radiation, one received chemotherapy during pregnancy, and the last patient refused treatment. All patients were followed every 3 months for evidence of recurrence. All suspicious local and distant metastases were biopsied. Radiation and/or chemotherapy were employed in the treatment of all recurrences.

We determined the significance of various clinical factors that may be responsible for survival. Among the 34 women treated for small cell NE carcinoma of the cervix, 25 had early-stage (Stage I–IIA) and 9 had advanced stage (Stage IIB–IVB) disease. Women with early-stage disease had median survival rates of 31 months compared with 10 months in the advanced stage group (P = 0.002). Of the patients with early-stage disease, the estimated 3-year and 5-year survival rates were 44.2% and 31.6%, respectively. All advanced stage patients died of their disease by 3 years (Fig. 1).

Figure 1.

Kaplan–Meier survival analysis based on International Federation of Gynecology and Obstetrics (FIGO) stage.

Eight cases showed an additional component of adenocarcinoma intimately admixed with small cell carcinoma. Immunohistochemical stains for cytokeratins were positive in seven of nine cases stained, epithelial membrane antigen (EMA) in two of two cases, neuron-specific enolase in six of six cases, chromogranin in six of nine cases, S-100 in one case, serotonin in one case, vimentin in neither of two, Leu7 in no cases, and leukocyte common antigen in no cases. At least one neuroendocrine marker (neuron-specific enolase, chromogranin, S-100, Leu7, or serotonin) was positive in each case stained.4, 6 The presence of dense-core granules, confirming neuroendocrine differentiation, was demonstrated on all electron microscopies performed. Pathologic prognostic factors including tumor size and lymph node involvement were analyzed (Table 2).

Table 2. Pathologic Characteristics
Variables%
Tumor size 
 ≤2 cm29% (10/34)
 >2 cm71% (24/34)
Tumor homology 
 Pure76% (26/34)
 Mixed24% (8/34)
Lymphovascular space invasion68% (17/25)
Lymph node involvement 
 Pelvic55% (12/22)
 Paraaortic70% (7/10)

Of the 23 women with early-stage disease who underwent hysterectomies, patients with tumor size >2 cm had median survival of 14 months compared with 155 months in those with <2 cm (P = 0.02). In fact, 6 of 8 early-stage patients with small tumors (<2 cm) and negative margins on the hysterectomy specimen were cured with a median follow-up of 117 months. Of the 6 patients who were cured of their disease, the majority (67%) underwent postoperative radiotherapy with or without chemotherapy. Of the 13 patients who underwent surgery with tumors >2 cm, 11 patients died of tumor progression, 1 patient was alive with disease at the time of last follow-up, and the remaining patient had no evidence of disease after receiving adjuvant radiotherapy.

Other pathologic variables such as lymph node metastasis, tumor homology, margin status, and lymphovascular space invasion also were evaluated. Of the 23 patients who underwent a pelvic and/or paraaortic lymph node dissection, 57% (13 of 23) had lymph node invasion that included 80% (12 of 15) of tumors >2 cm and 13% (1 of 8) of tumors <2 cm. Of the 13 patients with lymph node involvement, only 1 patient with a tumor <2 cm was alive without any evidence of disease at the time of last follow-up. However, in this patient population, lymph node metastasis was not found to be a predictor for survival (P = 0.46). Of the 23 patients with early-stage disease who underwent hysterectomies, 74% (17 of 23) had negative margins with a median survival of 58 months compared with 13 months for those with positive margins (P = 0.02). Moreover, 76% (26 of 34) of tumors had a pure histologic pattern with small cell NE carcinoma, and 24% (8 of 34) had a mixed histologic pattern comprised of glandular and/or squamoid features. Pure rather than a mixed histologic pattern was found to be a poor prognostic factor for survival (P = 0.04). Although lymphovascular space invasion was found in 68% of the tumors, it did not appear to have any prognostic significance in this patient population (P = 0.29) (Tables 2, 3).

Table 3. Univariate Analysis of Survival Based on Clinical and Pathologic Factors
VariablesMedian survival (mos)P value
  1. NR*: median survival not reached; HRT: hormone replacement therapy.

Smoking  
 Smokers11.6 
 Nonsmokers19.30.044
Stage  
 I–IIA30.6 
 IIB–IV9.80.002
Tumor size (early stage disease)  
 ≤2 cmNR* 
 >2 cm14.10.017
Tumor homology  
 MixedNR* 
 Pure14.10.044
Margin status  
 Positive12.8 
 Negative57.70.016
Age (yrs)  
 ≤4314.1 
 >4315.80.79
HRT  
 Users7.6 
 Nonusers19.30.139
Adjuvant chemotherapy  
 Yes15.8 
 No15.10.556
Adjuvant radiotherapy  
 Yes15.1 
 No30.80.198
Lymph node involvement  
 Yes14.1 
 No30.80.46

Demographic factors that may be important in the survival of patients with small cell NE carcinoma of cervix were evaluated. Women who were current smokers had a significantly poorer survival rate compared with nonsmokers (P = 0.04) (Table 3; Fig. 2). Conversely, age (P = 0.79), race (P = 0.20), menopause status (P = 0.96), and hormone replacement therapy (P = 0.13) were not found to be important prognostic factors.

Figure 2.

Kaplan–Meier survival analysis based on smoking.

To examine the variables identified as important in univariate analyses further, a multivariate analysis was performed. Smoking remained as a significant independent poor prognostic factor for survival in early-stage disease. Furthermore, stage of disease was an important prognostic factor in the overall population. However, other factors such as tumor size, homology, and margin status were not found to be significant independent predictors for survival (Table 4).

Table 4. Multivariate Analysis of Survival Based on Clinical and Pathologic Factors
VariablesHazard ratioP value
Early stage (I–IIA)  
 Smoking2.080.037
 Tumor size >2 cm1.920.055
 Negative margins0.890.373
 Mixed histology0.290.771
All stages  
 Advanced stage (IIB–IV)2.740.006
 Smoking1.820.069
 Mixed histology0.930.352

Adjuvant treatment with radiation with or without chemotherapy for patients with early-stage disease was also investigated. Women who received platinum-combination chemotherapy had a median survival rate of 16 months compared with 15 months in those without adjuvant chemotherapy (P = 0.56). Furthermore, patients with early-stage disease who underwent adjuvant radiotherapy had a median survival of 15 months compared with 31 months in those without radiotherapy (P = 0.20). Moreover, of the 3 patients with early stage disease who underwent adjuvant chemotherapy combined with radiation, 2 patients died of disease 10 months and 13 months, respectively, after their diagnosis and the remaining patient was alive without disease at the time of last follow-up. Because of the limited number of patients in this series, a significant survival advantage in those who received postoperative radiation and/or chemotherapy could not be demonstrated.

Thirteen patients (38%) had disease progression despite multimodality therapy. Fifteen women (44%) had disease recurrence, which included 3 with local and 12 with distant recurrence. Of the three patients who had local pelvic recurrences, two had Stage IB1 disease and one had Stage IIB disease. Two patients underwent pelvic radiotherapy with and without chemotherapy; at the time of last follow-up, one was free of disease and the other was alive with disease. The remaining patient died of progressive disease. Twelve patients recurred distantly, which included 9 with Stage IB1, 2 with Stage IIA, and 1 with Stage IIIA disease. The majority of these distant recurrences were located in the supraclavicular region. Two women had recurrences to the bone: one to the mandible and the other to the tibia. One other patient recurred to the breast. Only two patients with distant recurrences were alive at last follow-up without evidence of disease after radiation therapy. Of these two patients, one received chemoradiation for an isolated bone lesion and the other underwent surgical resection for recurrence to the breast; both remained disease free for 67 months and 108 months, respectively, after treatment for their recurrences. One woman presented with Stage IB2 disease at 17 weeks gestation and received primary chemotherapy with cisplatin. Because of disease progression, she received additional cycles of chemotherapy, which included vincristine, doxorubicin, and cyclophosphamide. After delivery at term, she received 40 centigrays of pelvic radiotherapy but died of progressive disease 2 months later.

DISCUSSION

Neuroendocrine carcinoma is an uncommon and aggressive type of small cell cervical cancer. Although 62% (21 of 34) of patients had Stage I disease, 82% (28 of 34) had either disease recurrence or progression. The 5-year estimated survival rate in early-stage patients was only 32%, and all of the advanced stage patients died. Previous reports have revealed that these patients uniformly have a dismal prognosis despite various treatment modalities.7, 8

Because small cell carcinoma of the cervix occurs infrequently, it is difficult to perform a randomized, controlled clinical trial to determine optimal therapy. This current study analyzed a large series of patients diagnosed with small cell carcinoma of the cervix from a single university practice, which included an update of a previous reported series.9 All diagnoses met specific light and electron microscopic criteria for NE small cell carcinoma of the cervix. The objective was to identify the clinical and pathologic factors that are responsible for survival of women with this aggressive tumor.

As confirmed in this series, FIGO staging criteria is important in this rare cervical malignancy. In early-stage disease, patients with small (<2 cm) tumors were found to have significantly better survival rates than those with large (>2 cm) tumors in univariate analysis. In fact, all these patients who were alive without evidence of disease at the time of last follow-up were diagnosed at an early stage with lesions <2 cm. Similarly, Sheets et al. showed that patients with tumors <2 cm had longer progression-free survival than patients with >2 cm lesions.9 In the current study, tumor size was found to have a marginal significance in multivariate analysis (P = 0.055). With more patients, tumor size may ultimately prove to be an important independent prognostic predictor for survival.

Radical surgery appears to be an important component in the treatment of early-stage, small cell NE carcinoma of the cervix. Only those with small tumors (<2 cm) amenable to surgery and negative margins after hysterectomy are long-term disease-free survivors. Although to our knowledge there currently are no studies published to date that have compared radical surgery alone with multimodality treatment in early-stage disease, many authors have recommended adjuvant chemotherapy due to the aggressive behavior of this tumor.10, 11 Similarly, Sevin et al. also concluded from their series of 12 patients that radical surgery is an important component in the multimodality treatment of small cell NE carcinoma of the cervix.12 However, women with large lesions (>4 cm) did poorly despite radical surgical treatment in this current series. Recently, Bermudez et al.13 have recommended neoadjuvant chemotherapy containing vincristine, bleomycin, and platinum for patients with large lesions >4 cm. Based on his series of 13 patients who received neoadjuvant chemotherapy, it appears that preoperative chemotherapy may be a useful therapeutic tool to enhance the resectability of the large tumors to improve outcome.13

It is interesting to note that smoking was a significant independent prognostic factor for survival. To our knowledge, this is one of the first studies to identify this association. Smoking may be a cofactor in the carcinogenesis pathway in small cell NE carcinoma of the cervix. Furthermore, smoking-related tumors reportedly are correlated with more aggressive tumor biology. Similarly, Videtic et al. demonstrated that the nonsmokers had significantly better overall survival rates compared with smokers during concurrent chemoradiation treatment for small cell lung carcinoma.14 Smoking and human papillomavirus (HPV) infections are clearly interrelated based on epidemiologic studies with squamous cell carcinoma of the cervix.15–18 A recent study by Deacon et al. showed that smoking and HPV infections were independent factors responsible for cervical neoplasia. Furthermore, they demonstrated a dose-response effect of smoking on cervical neoplasia using HPV-infected nonsmokers as controls and concluded that smoking may behave synergistically with HPV to induce cervical neoplasia.17 In addition, studies by Stoler et al.2 and Wolber et al.19 demonstrated that small cell carcinoma of the cervix was strongly correlated with HPV infection. Both studies showed that > 85% of the patient population were infected with high-risk HPV types (HPV-16 or HPV-18). Moreover, Abeler et al. used in situ hybridization to demonstrate that only high-risk HPV types (HPV-16 or HPV-18) rather than low-risk types (HPV-6 or HPV-11) were correlated with small cell NE tumors of the cervix.3

In the univariate analysis, patients with tumors involving a mixed histologic cell type such as squamous cell or adenocarcinoma have significantly better survival rates compared with those with pure small cell tumors. Similarly, Silva et al. also discovered that other histologic cell types admixed with small cell tumor confers a better prognosis.20 In their series, 34% of patients with small cell NE tumors mixed with squamous cell or adenocarcinoma histology were alive without disease compared with only 8% of those with pure small cell carcinoma. Conversely, some authors have not found an association between mixed histologic pattern and improved survival.11, 13 In multivariate analysis of this current study, tumor homology did not prove to be an independent predictor for survival. Of the patients with early-stage disease, the majority (87%) with pure small cell histology had tumors >2 cm, whereas < 50% of the women with small lesions <2 cm had a pure histologic pattern. Thus, tumor size and tumor homology appear to be correlated in this patient population.

Fifteen patients with poor prognostic factors on their hysterectomy specimen underwent postoperative radiotherapy with or without chemotherapy. Of these patients, nine women died of their disease secondary to distant and local recurrences, emphasizing the relative radioresistance and chemoresistance of this tumor. In this small subgroup of patients who underwent adjuvant treatment, there was no apparent survival benefit in those who underwent radiation with or without chemotherapy. However, it is interesting to note that patients who received adjuvant therapy were those with large lesions and other poor pathologic risk factors. Recently, Chang et al.11 treated 23 early stage patients with postoperative chemotherapy: 14 received vincristine (V), doxorubicin (A), and cyclophosphamide (C) alternating with cisplatin (P) and etoposide (E); and 8 received cisplatin (P), vinblastine (V), and bleomycin (B). The 5-year survival was 68% for patients who received the VAC/PE regimen compared with 33% for those treated with the VBP combination (P = 0.0078). In addition, Boruta et al. performed a multivariate analysis evaluating only patients with early-stage NE cervical carcinoma treated with adjuvant chemotherapy with or without radiation. They found an improved survival for patients treated with chemotherapy containing VAC or PE versus other regimens.10

These chemotherapeutic regimens are similar to those in the treatment of small cell NE carcinomas of the lung. In light of the pathologic and behavioral similarities between these tumors, it appears reasonable to explore active regimens derived from clinical trials of small cell NE lung carcinoma to obtain insights into the treatment of a comparable tumor originating from the cervix. Unlike patients with small cell lung carcinoma, women with small cell cervical carcinoma appear to have a significantly better cure rate, particularly those patients with early-stage disease.21–23 However, NE cervical carcinomas appear to be related to HPV. Thus, it remains to be determined whether this approach will have any survival benefit in patients with small cell NE cervical carcinoma.

The results of the current study demonstrated that radical hysterectomy is an effective therapeutic option in patients with early-stage disease. Only patients with early-stage with small tumors amenable to surgery and negative margins were long-term disease-free survivors in this series. Furthermore, smoking in patients with early-stage and advanced stage disease were considered independent poor prognostic factors for survival. Lastly, the role of primary or postoperative chemoradiation is unclear and yields uniformly poor results, particularly with advanced lesions. An algorithm for the treatment of small cell NE cervical carcinoma based on the results of the current study and a review of the literature is provided in Figure 3. Clearly, the optimal therapeutic regimen for patients with advanced disease remains to be determined. Although multimodality therapy ultimately may prove to enhance survival, it was not evident based on the limited number of patients in the current series. Thus, a multicenter randomized controlled trial is warranted.

Figure 3.

Management scheme for neuroendocrine cervical carcinoma.

Ancillary