Radiation Therapy for Primary Squamous Cell Carcinoma of the Vagina: Stanford University Experience
Nick M. Spirtos, M.D., B.P. Doshi, M.D., Daniel S. Kapp, Ph.D., and Nelson Teng, Ph.D., M.D.
Department of Gynecology and Obstetrics, Division of Gynecologic Oncology, and Department of Therapeutic Radiology, Stanford University Medical Center, Stanford, California 9430
A retrospective analysis of 38 patients with primary squamous cell carcinoma of the vagina seen at Stanford University Medical Center from 1958 to 1984 was undertaken. Patients were analyzed with regard to symptoms, stage, treatment techniques, survival, patterns of failure, and complications. Eighteen patients were classified as FIGO Stage I, 5 as Stage II, 10 as Stage III, and 5 as Stage IV. The 5-year disease-free survival was 94% in Stage I, 80% in Stage II, 50% in Stage III, and 0% in Stage IV. Five patients (13%) had eight major complications secondary to treatment. Only 2 of 23 patients with Stage I or Stage II disease developed a recurrence. There was a significant correlation between dose and response in patients treated with radiotherapy.
Carcinoma of the vagina is an uncommon gynecologic malignancy with a
reported incidence of 1-2% of all gynecologic malignancies [1-5]. Radiotherapy
is the preferred mode of therapy since most patients are elderly and rarely
good surgical candidates. Additionally, the proximity of the bladder and
rectum often mandates a radical surgical approach with urinary and/or
fecal diversion in order to secure adequate margins. Historically, the
5-year survival rates for patients with this disease have been very low
. However, newer radiotherapeutic equipment and better techniques have
resulted in improved survival as documented in more recent literature
Materials and Methods
The hospital and department records of 38 patients having primary squamous
cell carcinoma of the vagina seen and treated at Stanford University Medical
Center from May 1959 to July 1984 were reviewed. Patients with carcinoma
in situ of the vagina and lesions metastatic to the vagina were
excluded. Vaginal cancer in patients with a history of cervical cancer
was classified as primary only if more than 8 years had elapsed between
the initial diagnosis of cervical cancer and the occurrence of the invasive
Table 1. Treatment Techniques for Primary Invasive
Note: EBR, external-beam radiation: Ortho, orthovoltage boost;
Patients ranged in age from 15 to 99 years. Eighty-two percent of the patients were over 50 years of age. Thirty-three of the 38 patients (87%) were symptomatic. Vaginal bleeding and dysuria were the most common symptoms (Table 2). Thirty-two of the 38 patients had abnormal vaginal cytology. In 30 cases findings consistent with dysplasia were present. The cells were dyskeratotic, had diminution of the cytoplasmic content, and/or distorted nuclear membranes. In two cases the cellular specimen contained squamous cells with marked atypia. One patient had normal cytology. In 5 cases the results of the cytologic interpretations were not available. These 5 patients were all symptomatic. Twenty-nine of the 32 patients with abnormal cytology were symptomatic as was the 1 patient with normal cytology. Thus, 3 patients having abnormal cytology were asymptomatic.
All patients underwent clinical staging. Eighteen patients were Stage I, 5 were Stage II, 10 were Stage III, and 5 were Stage IV. Of the 5 patients with Stage IV disease, 2 were IVa and 3 were IVb.
The actuarial 5-year survival rate was 94% for patients in Stage I, 80% for Stage II, 50% for Stage III, and 0% for Stage IV. Table 3 compares the 5-year survival rate of primary squamous cell carcinoma in the present series with that in the literature.
The Kaplan-Meier curves for life-table analysis of the 38 cases of squamous cell carcinoma of the vagina are shown in Fig.1. Stepwise regression utilizing Cox regression models is summarized in Table 4. Earlier stage and higher radiation dose had a positive influence on survival.
Four patients had previously undergone radiation therapy for genital tract cancers. Two patients had squamous cell cancers of the cervix and 2 patients had adenocarcinomas of the endometrium. The shortest time interval from the previous irradiation to the diagnosis of vaginal cancer was 15 years.
Table 2. Presenting Symptoms of Patients with Vaginal Carcinoma
Only 1 of 18 patients with Stage I and 1 of 5 patients with Stage II disease recurred. The patient with recurrent Stage I squamous cell carcinoma was initially treated with external-beam radiation (5000 rad) and brachytherapy using a vaginal candle (2000 rad to the tumor surface). Nine months after completion of therapy she underwent total pelvic exenteration for central recurrence. The patient with recurrent Stage II disease presented 3 years after the completion of therapy (6100 rad) complaining of shortness of breath. Pulmonary metastases were documented and the patient died shortly thereafter. Five out of the 10 patients with Stage III disease died. Four of these patients had large bulky disease and were treated with external-beam therapy alone which proved insufficient as a means to control the disease. Table 5 shows the relationship between radiotherapy dosage and recurrence in patients with vaginal carcinoma. In patients with Stage III disease, 4 of 5 treatment failures occurred in cases when the total radiation dose was less than 7500 rad. All 5 patients with Stage IV disease died.
Five patients developed eight major complications necessitating hospital admission. All complications were sequelae attributable to radiotherapy. One patient with Stage I disease developed rectal stricture requiring colostomy and subsequently developed partial intestinal obstruction requiring laparotomy. One patient with Stage II disease developed a rectovaginal fistula necessitating diversion via sigmoid colostomy. She subsequently developed fibrosis of the ileum and underwent a segmental resection with an end-to-end reanastamosis. Her post-operative course was further complicated by Pseudomonas pneumonia which ultimately proved fatal. Two Stage II patients developed major complications: one developed a rectal ulcer requiring colostomy and the other developed a urethrovaginal fistula, necessitating urinary diversion with an ileal conduit. One patient with Stage IV cancer developed a rectovaginal fistula and was treated with a sigmoid colostomy.
Table 3. 5-Year Survival Rates by FIGO Stage of Primary
Carcinoma of the vagina is a rare but easily detectable disease. The importance of thorough vaginal examination and Pap smear cannot be over emphasized since carcinoma detected at an early stage has a high cure rate. In the present series, cytologic sampling was performed in all 38 patients. The results of 33 were available and 32 were abnormal and 1 normal as previously mentioned. Leaving aside the 5 patients whose cytologic interpretation is unknown, there still remains 33 patients whose cytology results and clinical history are evaluable.
Interestingly, all but 3 of these patients were symptomatic. The single patient with a known normal smear was symptomatic and had a vaginal mass present on physical examination. Three asymptomatic patients, then, underwent evaluation and the diagnosis of sqamous cell cancer of the vagina made as a result of an abnormal Pap smear. All had Stage I disease and are alive at 5 years. Most commonly evaluation of symptoms directly related to the patients' cancer lead to the diagnosis of vaginal cancer. However, cytologic evaluation did lead to the diagnosis of three Stage I cancers (17%) which are, by all reports, the most curable. Thus, the role of cytologic screening cannot be minimalized in squamous cell cancer of the vagina. It is essential that these simple screening tools be utilized so as to detect this disease in its early, more curable stages.
Table 4. Primary Squamous Cell Carcinoma of the Vagina: Stepwise
(a) Total dose in rads to vaginal mucosa.
The stage of the disease is the most important prognostic factor in patients with vaginal cancer (P = .001). In our experience, patients with Stage I disease had a 5-year survival rate of 94% (17/18). Results reported in many other series pale by comparison [6,10,18,19]. However, Perez 1, Pempree and Amornmarn 2, and Gallup  have reported similar success in treating Stage I cancer of the vagina with 5-year survivals of 95% (36/38),83% (5/6), and 100% (4/4), respectively. Perez  reported 36/38 patients with Stage I disease were alive without evidence of disease at 5 years. However, it should be noted that 11/115 total patients in that series did not have squamous cell cancer of the vagina. Thus we are unable to assume that all 38 Stage I patients had squamous cell histology or that all 36 survivors had squamous cell and not other histologic variant. It should be noted that besides reporting comparable 5-year survival data, Perez  treated patients similarly to ours, especially in terms of radiation dosage and method of delivery. Perez reported 33/38 (87%) patients with Stage I disease received either brachytherapy alone or in combination with external-beam radiation. Fifteen of 18 (83%) similarly staged patients in our series were treated in a likewise manner. Pempree and Amornmarn  reported that 5 of 6 patients with Stage I disease survived at least 5 years. However, it is not clear how many had squamous cell cancers. Overall, 60/64 (94%) patients had squamous cell lesions, but this figure is not broken down by stage.
For reasons not unlike the one just mentioned, comparing the results of one series with another is a difficult task. Due to the relative rare occurrence of vaginal cancer most series must retrospectively examine cases spanning two to three decades. Undoubtedly in this situation, standardization of treatment is nearly impossible within any one series much less between two. Retrieval of exact dosimetry and means of delivery is also difficult to obtain in retrospective studies. Finally, different techniques of statistical analysis makes comparison of results difficult. For example, Perez , Houghton and Iverson , Rubin et al. , and the present series use actuarial 5-year survival whereas Pempree reported absolute 5-year survival figures.
Table 5. Radiotherapy Dosage with Respect to Recurrence in
Note. Numbers in parentheses denote patients with recurrences.
Although the number of patients having Stage II disease in our series was small, the results were similar to other reports in the literature. Again, comparing survival data from different series is difficult both because of statistical differences as well as the inclusion of different histologic cell types in the various reports. Table 6 collates the 5-year survival data from series where the treatment method and dose were known and squamous cell lesions could be separated from other cell types or was the only cell type analyzed.
Some authors [2,9,20] have reported less success treating Stage IIB compared to Stage IIA disease. (See  regarding modification of FIGO staging by subdividing Stage II into IIA and IIB.) On the other hand, Puthawala et al.  reported similar survival data for patients with Stage IIA or IIB disease. It should be noted that six of his patients with IIA disease received approximately 8000 rad total dose whereas patients with IIB disease received approximately 10,000. Our results seem to substantiate Puthawala's experience. Three of our five patients had parametrial involvement and two of these three patients are free of disease at 5 years. The only patients who died of disease had positive para-aortic nodes. All of the Stage II patients in our series received at least 7500 rad total dose delivered via combination external-beam radiation and brachytherapy and four of five were alive without evidence of disease at 5 years. The poor response when external-beam radiation was used alone is noteworthy (Table 6). We are unable to speak to the effectiveness of brachytherapy alone compared to brachytherapy combined with external-beam radiation since no randomized study comparing the two methods, much less one that controlled for tumor volume, has been performed.
Table 6. Stage II: 5-Year Survival and Treatment Methods
In light of these data, we believe radiotherapy is an excellent method of treatment for patients with Stage I and II vaginal carcinoma. Local tumor control and significant 5-year survival can and have been achieved by giving tumorcidal doses of 7500-10,000 rad through the careful integration of external-beam and brachytherapy techniques. For tumor depths 5mm or greater, a vaginal candle should not be utilized due to poor penetration. Instead, external-beam treatment followed by interstitial implants should be used. Theoretically, the higher dose achieved using this technique increases the risk of complications to the bladder and rectum. However, this can be minimized by judicious integration of shaped externalbeam fields followed by brachytherapy. The importance of brachytherapy in the treatment of vaginal cancer cannot be overemphasized and has been well documented in the literature . Brachytherapy alone can be used in the treatment of selected patients with Stage I disease with excellent local control rates. This is demonstrated in the present series as well as others in the literarature. All seven patients (100%) with Stage I disease treated with brachytherapy alone at Stanford were alive without evidence of disease 5 years after therapy. Similar survival data has been reported by Pempree 2/2 (100%) ; Perez 12/13 (92%) ; Nori 4/5 (80%) ; and Rubin 2/3 (67%) .
Five of our 10 patients with Stage III disease survived 5 years and this again compares favorable with other investigators' reported 5-year survivals. Notable in this group of patients is that all patients treated by brachytherapy alone or external-beam radiation alone died of disease. This experience parallels others reported in the literature. Therefore, most authors recommend an integrated approach using external-beam radiation in conjunction with brachytherapy (Table 7). The results presented in this series support this approach. Survival in patients with Stage III disease could be related directly to the radiation dose received. Four of the five patients receiving more than 7500 rad survived whereas only one of five patients receiving less than 7500 rad survive. Review of the recent literature regarding the relationship of dose to survival is difficult to analyze for several reasons, the most important being that tumor volume, treatment methods, and doses, as well as lymph node status, are not always noted. Even in those reports where the method and dose data are available, tumor volume and lymph node status most often goes unmentioned, and thus drawing a singular conclusion is difficult. Results reported by Pempree and Amornmarn  and Perez , in addition to ours, lead to the conclusion that survival is affected by the dose delivered. Most patients in these series received greater than 7500 rad and the pooled 5-year survival is 40% (18 of 44). Peters et al.  also reported results similar to ours: five of six patients receiving more than 6500 rad were alive 5 years after treatment whereas only one of six patients with Stage III or IV disease receiving less than 6500 rad survived 5 years. Unfortunately, Peters et al. analyzed patients with Stage III and IV disease as one group when, in fact, our results, as well as others, would indicate there is a real difference in the response between the two stages.
Table 7. 5-Year Survival of Patients with Stage III Disease
Interestingly, Puthawala et al.  and MacNaught et al.  have reported less impressive results when treating patients with reportedly similar doses and methods. Pooling the survival data in their reports reveals only 27% (7/26) patients with Stage III squamous cell cancer were alive without evidence of disease 5 years following treatment. Puthawala et al. and MacNaught et al. treated all patients with more than 7500 rad. It is not obvious why the survival in these two reports is one-third that reported in other series. One variable which has received little attention that might explain the wide range of survival rates reported is the lymph node status of patients with advanced stage vaginal cancer. Although none of our patients with Stage III disease underwent staging laparotomy, 7 of 10 did have lymphangiograms (LAGs) performed. Four were positive and 3 were negative. Three of the 4 patients with positive LAG's died of disease. Unfortunately, only 1 was treated with more than 7500 rad. One of 3 patients with a negative LAG died of disease. This patient received more than 7500 rad. Though the numbers are small and the limitations of lymphangiogram is so poor. Nine of the 14 patients with Stage III disease reported by Rubin et al.  had lymphangiograms. Seven of 9 were positive. Unfortunately, it is indiscernable how many of the patients with positive LAGs survived 5 years. Rubin's findings regarding the number of Stage III patients with positive LAGs are similar to ours. The significance of these findings is yet to be determined. A prospective clinical-pathologic study is necessary to determine the importance of lymph node metastases in advanced vaginal cancer.
We were unsuccessful treating patients with Stage IV disease. Again our results are similar to others reported in the literature [2,9,17,22]. There are few reported survivors treated primarily by radiation therapy regardless of dose or method of delivery. All 5 of the patients treated at Stanford University Hospital had widely metastatic disease at the time of their demise. Other authors concur that even when local control is achieved most patients with Stage IV disease succumb to metastatic cancer (Table 8). These results would indicate that treatment of Stage IV disease most likely requires an integrated approach utilizing systemic chemotherapy as well as radiotherapy. Unfortunately, the role of systemic chemotherapy has yet to be adequately defined for patients with advanced squamous cell cancer of the vagina. Thigpen et al.  reported that only 1 of 16 patients with advanced or recurrent squamous cell cancer of the vagina responded to cisplatin (50 mg/m2) and this patient had not received any other prior therapy. As with patients with advanced or recurrent squamous cell cancer of the cervix, few respond to chemotherapy following other treatment modalities.
Table 8. Stage IV Site of Failure
Five of 38 patients (13%) suffered major complications as a result of
irradiation. Three of 5 had advanced disease (Stage III or IV), 1 patient
had Stage II, and 1 patient had Stage I disease. These results are similar
to those of Houghton 8/67 (12%) : Puthawala 4/27 (14%) ; Rubin
8/61 (13%) ; and Gallup 4/28 (14%) . Four of the 5 patients suffering
major complications secondary to irradiation are free of disease at 5
years. Only the patient having Stage IV disease and complications has
died. Each of these 5 patients received a minimum of 7600 rad to the vaginal
surface (7600-11,200). We had no complications in those patients receiving
less than 7500 rad. However, 6 of 13 patients (Stage I-III) receiving
less than 7500 rad did recur and die as opposed to only 2 of 22 (Stage
I-III) receiving greater than 7500 rad. In fact, the rate of recurrence,
regardless of stage, in patients who received less than 7500 rad was greater
than the rate of recurrence in those receiving greater than 7500 rad to
the tumor volume (P = .034). All patients with recurrence ultimately
died of disease. This fact certainly emphasizes that patients with Stages
I through III disease should be considered for curative treatment and
receive at least 7500 rad. Thus, the risk of the larger dose to our way
of thinking is well worth taking.
1. Rutledge, F. Cancer of the vagina. Amer. J. Obstet. Gynecol. 97, 635