66_05

Treatment of Recurrent Carcinoma of the Cervix

After Definitive Irradiation

Reirradiation of previously irradiated patients must be undertaken with extreme caution. It is very important to analyze the techniques used in the initial treatment (beam energy, volume, doses delivered with external or intracavitary irradiation). Also, the period of time between the two treatments must be taken into consideration because it is postulated that some repair of the initial damage may take place in the interval. However, it is foolhardy to assume that previously irradiated tissues will have the same tolerance as newly irradiated tissues. In general, external irradiation for recurrent tumor is given to limited volumes (40 to 45 Gy, 1.8-Gy tumor dose per fraction, preferentially using lateral portals). Occasionally, intracavitary or interstitial irradiation can be used to treat relatively circumscribed recurrences.

Sommers et al. (566) described the results of retreatment in 376 patients with recurrent carcinoma of the uterine cervix. Ninety-one patients received irradiation, mostly external (86.8%), occasionally combined with brachytherapy (7.7%) to control bleeding of central recurrences; brachytherapy alone was administered in 5.5% of patients. The usual dose for recurrent pelvic masses was 40 to 45 Gy and for para-aortic lymph node metastases 45 to 50 Gy in 5 weeks. Other metastatic sites were treated with 35 to 40 Gy in 3 to 4 weeks. Pelvic exenteration was attempted in 23 patients, only 10 of whom were deemed to be operable (43.5%), but it was completed in only seven. The probability of 5-year survival after treatment for recurrence was 30% with combined surgery and external irradiation, 12% with surgery alone, and 4% with external irradiation alone. The 5-year survival rate for 10 patients who underwent pelvic exenteration was 16%. Only 1% of the untreated patients survived 5 years. Six of 140 patients (4.3%) experienced grade 2 or 3 complications.

Selected patients with limited pelvic recurrences not fixed to the pelvic wall and without evidence of extrapelvic metastases can be potentially salvaged by radical hysterectomy or pelvic exenteration.

Coleman et al. (91) described results in 50 patients who underwent radical hysterectomy for persistent (18 patients) or recurrent (32 patients) cervical cancer after primary radiation therapy. Lymph node metastases were identified in 5/39 patients (13%) in whom the lymph nodes were evaluated. The 5- and 10-year survival rates were 72% and 60%, respectively.

In 65 patients on whom pelvic exenteration was carried out at Memorial Sloan-Kettering Cancer Center, the 5-year survival rate was 23%. The operative mortality rate was 9.2% (357). The authors pointed out that the significant mortality and morbidity associated with this procedure preclude its use as palliative therapy.

Urinary diversion, either by nephrostomy or ileal bladder, may be of palliative value in patients with either recurrent carcinoma in the pelvis or complications. It must be kept in mind that diversion may prolong life but runs the risk of denying a terminally ill patient with cancer the oblivion and insensibility of uremia.

Kastritis et al. (301) treated 200 patients with stage IV or recurrent cervix cancer with platinum-based chemotherapy; response rate was 43.5% in 142 with squamous-cell and 53.5% in 58 with nonsquamous tumors (p = 0.79). Median survival was 11.57 and 19 months, respectively.

Tinker et al. (613) treated 25 women for recurrent cervical cancer with carboplatin-paclitaxel and noted a 20% cure rate and 20% progression rate, with median survival of 21 months.

Brewer et al. (54), in 32 women all of whom had previous chemotherapy and 29 previous RT, used cisplatin and gemcitabine, with a progression rate of 22% and median time to progression 3.5 months.

Para-Aortic Lymph Node Recurrences

Isolated recurrences in the para-aortic nodes after pelvic irradiation have been described in about 3% of patients and some may be salvaged with aggressive therapy. The advent of IMRT and IGRT makes treatment easier, with less morbidity.

Kim et al. (319) treated 12 patients with isolated para-aortic lymph node metastasis with hyperfractionated RT (60 Gy in 1.2 Gy fractions twice a day) and concurrent cisplatin-paclitaxel. Fields extended from superior plate of T12 to lower plate of L5. Three-year survival was 19%. Grade 3 or 4 hematologic toxicity developed in two patients.

Singh et al. (560) detected 14 isolated para-aortic lymph node metastases in 816 patients previously treated with RT, who subsequently received RT to the para-aortic lymph node combined with concurrent chemotherapy. Seven patients survived 5 years.

Jhingran et al. (283), in a review of 1,955 patients treated with RT for cervix cancer, identified 120 with recurrent tumor above the pelvic fields. Initially, 10 had common iliac and five para-aortic node involvement. In 104 patients, recurrences were immediately adjacent to the upper borders of the RT fields. In 15 patients treated with curative intent for the para-aortic lymph node recurrence, 5 year survival was 25%.

After Previous Surgery

It is substantially easier to treat surgical recurrences with irradiation, which can salvage approximately 50% of patients with localized pelvic recurrences after surgery alone. A combination of whole pelvis external irradiation (20 to 40 Gy), depending on the volume of tumor, and additional parametrial dose with midline shielding for a total of 50 to 60 Gy are needed. In addition, one or two LDR intracavitary insertions (or equivalent HDR) that may cover either the vaginal vault or the entire vagina, depending on tumor volume, should be delivered. The total mucosal dose from the external and LDR intracavitary therapy can approach 140 Gy to the upper vagina and 95 Gy to the distal vagina without a high risk (248). It is extremely useful to combine these techniques with interstitial irradiation to boost the dose to residual. Doses of 20 to 35 Gy are administered with single, double-plane, or volume implants, depending on the extent of the tumor.

Larson et al. (354) observed 27 recurrences (11%) in 249 patients treated with radical hysterectomy and pelvic lymphadenectomy for stage IB carcinoma of the cervix; 17 (63%) had tumor recurrence in the pelvis or vulva; the other 10 patients had recurrences outside the pelvis. Eight of 15 patients (53%) treated with irradiation for an isolated recurrence in the pelvis or vulva were tumor free between 10 and 126 months after treatment of the recurrence (median, 48 months).

Ijaz et al. (268) reported on 50 patients treated with RT for an isolated pelvic recurrence of cervical carcinoma after radical hysterectomy; seven patients were treated with palliative intent using hypofractionated RT. The remaining 43 patients were treated with curative intent, 33 with RT only and 10 with cisplatin-based chemoirradiation. The overall 5-year survival rate was 33% for all 50 patients, 39% for the 43 patients treated with curative intent, and 25% for patients with isolated sidewall recurrences treated with curative intent. Three patients experienced late treatment complications.

Hille et al. (246) described results in 17 patients with recurrent cervix cancer (nine had a complete microscopically incomplete resection) treated with EBRT and BT to 50 to 65 Gy. The 5-year pelvic tumor control was 48% and RFS 24%.

P.1591

Intraoperative Irradiation

Intraoperative radiation therapy (IORT) has been used for treatment of not only locally advanced but recurrent carcinoma of the cervix, with 3-year survival rates of 8% to 21% as reported by Mahé et al. (393) and Garton et al. (174), and a 5-year survival rate of 33% in 14 patients described by Kinney et al. (324). Patient selection may have had an impact on the different results. Abe and Shibamoto (1) noted that central recurrences, particularly in nonirradiated patients, and resection of the gross recurrent tumor in irradiated patients improve the benefit from IORT. Significant toxicity included peripheral nerve injury and ureteral stenosis (with doses higher than 15 to 20 Gy).

IORT was used in 70 patients with pelvic recurrences in a European cooperative study (393). Complete tumor resection was carried out in 30 patients, partial in 37, and unspecified in 3. Sixty-five patients had electron beam therapy (12 to 25 MeV), with mean doses of 18 Gy (10 to 25 Gy) after gross complete resection and 19 Gy (10 to 30 Gy) after partial resection. The 3-year overall survival rate was 8%. Grade 2 or 3 toxicity was observed in 19/70 patients (27%), with 10 complications being related to IORT.

Martinez-Monge et al. (403) reported a study of IORT in 26 patients with recurrent gynecologic tumors, some relapsing after full-dose radiation therapy (group 1) or recurrent disease after surgery (group 2). Cervical carcinoma was the initial tumor site of involvement in 18 patients. Treatment consisted of maximal surgical resection and IORT (10 to 25 Gy) to high-risk areas. Patients not previously irradiated also received external-beam irradiation (with or without chemotherapy) before or after surgery. There was one IORT-related incidence of motor neuropathy. The local tumor control rates were 33% and 77%; the 4-year actuarial survival for group 1 was 7% and the 6-year actuarial survival rate for group 2 was 33%.

Chemotherapy

Chemotherapy is being more extensively used in bulky and advanced cervical cancer; some cytotoxic agents have shown encouraging efficacy in patients with advanced and recurrent cervical carcinoma (662).

Noteworthy, in a study of 41 patients who had weekly biopsies while receiving RT and chemotherapy for cervix cancer, increased tumor cell proliferation and accelerated repopulation was observed within 2 weeks from the initiation of therapy. Patients with a sustained yield and high S-phase fraction for two or more weeks were at increased risk for tumor progression (131).

Green et al. (189), in a search of medical databases for randomized trials of cervical cancer that compared RT with or without concurrent chemotherapy, identified 19 trials comprising 4,580 randomized patients, and they were the subjects of the meta-analysis. Concomitant chemotherapy and radiation improved tumor control and overall survival (RR 0.71; p <0.0001) and progression-free survival (RR 0.61; p <0.0001). The benefit was maximal in early stage (I and II) disease. The absolute survival benefit was 12%. Patients receiving chemoirradiation had a higher incidence of grade 3 or 4 hematologic and gastrointestinal toxicities.

Cisplatin is one of the most active cytotoxic agents in squamous-cell carcinoma of the uterine cervix (529). When cisplatin and irradiation are used concomitantly, substantial enhancement of cell killing is observe.

Coughlin and Richmond (98) and Douple (125) suggested two mechanisms for radiation enhancement by platinum: (a) in hypoxic or oxygenated cells, free radicals with altered binding of platinum to DNA are formed at the time of irradiation; and (b) interaction inhibits repair of sublethal damage.

Cisplatin has been combined with other cytotoxic agents. Long et al. (384) conducted a randomized study comparing methotrexate, vinblastine, doxorubicin, and cisplatin (MVAC) or cisplatin/topotecan or cisplatin alone in patients with advanced cervical cancer. The MVAC arm was closed after four deaths in 63 patients occurred. In 294 patients assigned to the other arms response rate was 27% for cisplatin/topotecan and 13% for cisplatin alone, with median survival of 9.4 and 6.5 months, respectively. Other trials with cisplatin will be discussed later in this chapter.

Paclitaxel, a natural product found initially in the bark of the western yew tree, produces depolymerization and irreversible bundling of tubulin in the cell. It has been shown to have a radiosensitizing effect.

Rose et al. (522) reported on a phase II study of 44 patients in which the starting dose was paclitaxel (135 mg/m2, maximum 170 mg/m2) infused over 24 hours, followed by cisplatin (75 mg/m2) every 21 days. Forty patients (90.9%) had received prior radiation therapy. A median of six courses of chemotherapy was given. Of the 41 assessable patients, five (12.2%) had a complete response and 14 (34.1%) had a partial response.

Vinorelbine is a semisynthetic derivative of vinblastine. In a phase II trial in patients with prior irradiation, an 18% response rate was observed (464).

A second phase II trial used the drug as neoadjuvant chemotherapy; in 42 patients, two complete and 17 partial responses (45%) were observed (345,422).

Irinotecan and topotecan are camptothecin derivatives whose cytotoxic mechanism is thought to target topoisomerase I (623).

An international phase II trial reported a similar 21% response rate in patients predominantly with prior irradiation (one complete and eight partial responses among 42 patients) (635).

Gemcitabine, a nucleoside analog, showed a 4.5% partial response and 36% stable disease in 22 patients (546).

In combination with cisplatin, it was evaluated in 32 women with previously treated cervix cancer (initial dose 800 mg/m2 days 1 and 8 then every 28 days); there were seven (22%) partial responses and 12 stable disease responses (54).

Nonrandomized Studies of Chemotherapy and Irradiation

Numerous preliminary reports have been published on results of neoadjuvant/concomitant use of cisplatin and 5-fluorouacil (5-FU), with or without mitomycin C, combined with irradiation to treat patients with locally advanced or recurrent carcinoma of the cervix (46).

Trials with Cisplatin, 5-Fluorouracil, or Both

Perez and Grigsby (477) reported on 58 patients with locally advanced carcinoma of the cervix treated with concurrent 5-FU/cisplatin and irradiation, and compared the results with 257 patients with similar stages treated with irradiation alone during the same period. Pelvic tumor control, disease-free, and cause-specific survival were comparable. The incidence of rectal and bladder fistula was 7% in the chemoirradiation group and 4% with irradiation alone (p = 0.61).

Park et al. (459) described results in 113 patients with high-risk invasive cervical carcinoma treated with cisplatin and 5-FU. For adenocarcinoma, doxorubicin (45 mg/m2 IV) was added. The patients subsequently received radiation therapy (not described in the publication). Compared with 77 patients treated with RT alone, in stage I or II tumors >4 cm, the 5-year survival rate was 78.3% with chemoirradiation and 48% with RT alone (p <0.01) and in stages III and IV 69.1% and 57.4%, respectively. Toxicity with combined chemoirradiation was not significantly enhanced compared with irradiation alone.

P.1592

Sardi et al. (544) reported results of three courses of cisplatin, vincristine, and bleomycin (days 1 to 3) at 10-day intervals combined with RT in 205 unselected patients with stage IB cervix cancer (tumors >2 cm) who were divided at random into two groups treated with surgery and RT or neoadjucvant chemotherapy, surgery, and irradiation. After 67 months, no difference in survival was seen in patients with tumors 2 to 4 cm in both groups (77% for control patients vs. 82% with neoadjuvant chemotherapy), but statistically significant differences were seen in bulky tumors (>4 cm), 61% versus 80% in favor of neoadjuvant chemotherapy.

Souhami et al. (574) treated 50 patients with bulky, locally advanced carcinoma of the cervix with a combination of weekly cisplatin (30 mg/m2) concurrent with RT. At 44 months, the actuarial survival rate was 65%, the total pelvic failure rate was 26%, and the distant metastasis rate was 24%. The incidence of late gastrointestinal toxicity was high, with 10 rectal ulcers (four colostomies required for severe bleeding), two rectovaginal fistulas, and two small bowel obstructions.

Park et al. (460) treated patients with stage I and II carcinoma of the cervix larger than 4 cm with RT alone, concurrent or sequential chemoradiation with cisplatin and 5-FU. The 30-month survival rates were 100% with concurrent chemoirradiation, 89.5% with sequential treatment, and 79.5% with irradiation alone (p <0.05).

Varia et al. (631) reported on GOG Protocol 125 that 87 patients with biopsy-confirmed para-aortic lymph nodes from cervical cancer were treated with extended-field irradiation (45 Gy in 1.5-Gy fractions) and higher doses to the pelvis (approximately 80 Gy to point A) in combination with 5-FU and cisplatin. The 3-year progression-free survival rate was 33%, and the overall survival rate was 39%. Grade 3 and 4 hematologic toxicity was noted in 13 patients (15%), chronic proctitis in three (3.5%), and four patients (4.6%) required surgery for rectal complications.

Lee et al. (366) treated 40 women with cervix cancer using 50 Gy EBRT and brachytherapy; in 25 three concurrent cycles of cisplatin/5-FU and in 15 six cycles of consolidation chemotherapy were given. There was no difference in 2-year survival between the two groups (98% to 100%). Grade 2 or greater hematologic toxicity was more frequent in the consolidation patients.

Grigsby et al. (197), in a prospective study of 65 patients with cervical cancer and node-negative on FDG-PET treated with RT alone (15 patients) or combined with concurrent weekly cisplatin (50 patients), noted a 5-year cause-specific survival of 78% and 74%, respectively. Severe complications included one RV fistula and one rectal stricture in the concurrent chemotherapy/RT group and one chemotherapy-related death.

Trials with Mitomycin C or Tirapazamine

Mitomycin C acts as an alkylating agent and inhibits DNA and RNA synthesis. Activation of mitomycin C is increased in hypoxic conditions, and thus it acts as a hypoxic radiosensitizer. Interstitial pneumonitis and pulmonary fibrosis are usually related to the dose of drug. Use of IV dexamethasone before administration of the drug may prevent pulmonary toxicity.

Christie et al. (83) described results in three groups of patients with stage IIB and III carcinoma of the cervix treated with pelvic irradiation and an intracavitary insertion combined with chemotherapy. Group A (64 patients) received 5-FU infusion during the first and last weeks of irradiation combined with mitomycin C (10 mg/m2 IV). Group B (29 patients) received 5-FU without mitomycin C, and group C (84 patients) received irradiation alone. With median follow-up of 7.2 years, the 5-year survival rates were 56%, 32%, and 36%, and the local tumor control 73%, 53%, and 50%, respectively. Toxicity was greater in group A (36% grade 3 and 4) compared with the 5-FU and irradiation group (14%) and the irradiation-alone group (20%).

Roberts et al. (513) reported on a trial in which 160 patients with locally advanced cervical cancer were randomized to receive RT alone (82 patients) or RT with concomitant mitomycin C (78 patients). The 4-year actuarial survival was 72% and 56%, respectively (p = 0.13), and the local recurrence-free survival rate was 78% and 63%, respectively (p = 0.11). There were no treatment-related deaths. No excess in nonhematologic toxicity has been observed with combined mitomycin C and irradiation.

Tirapazamine, a radiation sensitizer, with selective cytotoxic effect on hypoxic cells, was combined with cisplatin in 56 patients with recurrent or metastatic cancer. After six cycles given every 21 days, four complete and 13 partial responses were noted. Overall 6-month survival was 56%. Better response was seen in patients who had not received radiosensitizing chemotherapy previously (562).

Besides the usual hematologic and pelvic toxicity described in many of these studies with chemoradiation, Wun et al. (666), in a retrospective analysis of 75 patients with gynecological cancer who received erythropoietin and chemotherapy/RT, noted that 17 had upper or lower extremity thrombosis, in contrast to 2/72 who did not receive erythropoietin. Noteworthy, Anders et al. (17), in a review of the literature, reported 6/128 patients (4.7%) treated with chemptherapy/RT without erythropoietin developed grade 4 or 5 thrombosis toxicity.

Intra-Arterial Chemotherapy

Intra-arterial infusion of chemotherapeutic agents in cervical carcinoma was used for some years based on the distinct arterial supply to the tumor-bearing area. Unfortunately, the responses have been uncommon and short, and the toxicity and complication rates have been significant (423).

Onishi et al. (454) evaluated intra-arterial cisplatin through catheters inserted into both internal iliac arteries in cervix carcinoma. Patients were randomized into a concurrent intra-arterial infusion of platinum with RT (18 patients) or RT alone (15 patients). Five-year overall survival rates were 44.4% and 50%, respectively. In the group receiving intra-arterial infusion, grade 3 or 4 late bowel complications were seen in 44% and grade 3 or 4 myelosuppression in 33%, significantly more than in the RT group.

Randomized Studies of Chemotherapy and Irradiation

Hydroxyurea

The GOG carried out a trial of irradiation with either concomitant hydroxyurea (HOU) or a placebo in patients with stage IIIB or IVA cervix cancer (263).

The study was criticized because patients were not surgically staged, half of the 190 patients were not evaluable, and radiation doses were low (478).

Piver et al. (487) published an update of a study of 130 patients (13 with para-aortic lymph node metastasis), 75 of whom were surgically staged. Of 66 patients who underwent surgical staging, 33 received hydroxyurea and 33 a placebo in combination with irradiation. Of the patients who did not have surgical staging, 27 received hydroxyurea and 37 received placebo. The 2-year survival was higher in the HOU group.

In another larger randomized trial by the GOG reported by Stehman et al. (579), 296 surgically staged patients with stage IIB to IVA disease and negative para-aortic nodes were randomized to irradiation plus either hydroxyurea (139 patients) or misonidazole (157 patients). Survival was not statistically different between the regimens, with 33.8% deaths in the hydroxyurea and 38.9% deaths in the misonidazole groups (p = 0.25).

P.1593

Failure limited to the pelvis occurred in 18% of patients in the hydroxyurea group and 23.6% in the misonidazole group.

Noteworthy, in a randomized RTOG trial of patients with stage III disease, Leibel et al. (371) and Overgaard et al. (456) reported lower survival in patients receiving misonidazole than in the patients treated with irradiation alone.

Symonds et al. (592), in a review of seven randomized trials, found no evidence to support the use of HOU with RT in cervix cancer.

Cisplatin

Tattersall et al. (600) published a report of 71 patients with stage IB and IIA carcinoma of the cervix treated by radical hysterectomy who had metastatic pelvic lymph nodes and were entered into a randomized trial comparing pelvic RT alone or combined with three cycles of cisplatin (50 mg/m2), vinblastine (4 mg/m2), and bleomycin (15 mg), followed by the same pelvic irradiation. There was no difference in survival in the two groups.

Later, Tattersall et al. (598) reported on a randomized trial of 260 patients with stage IIB to IVA cervical cancer, 131 treated with pelvic irradiation (45 to 55 Gy followed by 30 to 35 Gy LDR brachytherapy), and 129 receiving chemotherapy (cisplatin and epirubicin at 3-week intervals for three cycles) followed by similar pelvic irradiation. Patients who received chemotherapy had a significantly higher pelvic failure rate (29% vs. 19%; p <0.003) and inferior 3-year disease-free survival (40% vs. 50%) compared with those treated with RT alone (p = 0.02).

Souhami et al. (573) randomized 107 patients with stage IIIB carcinoma of the cervix to treatment with irradiation alone or combined with bleomycin, vincristine, mitomycin, and cisplatin (BOMP). The overall 5-year survival rate for the neoadjuvant-treated patients was 23% in contrast to 39% for those treated with irradiation alone (p = 0.02). Locoregional and distant failure rates were similar in both groups.

Chiara et al. (80) described a randomized study of patients with advanced carcinoma of the cervix (stages IIB and III) in which 46 patients were treated with irradiation alone or in combination with cisplatin. The actuarial 3-year progression-free survival rates were 72.4% and 59.3%, respectively.

Kumar et al. (342) reported a randomized trial in which 94 patients with carcinoma of the cervix were treated with chemotherapy (two cycles of bleomycin, ifosfamide-mesna, and cisplatin) followed by RT, and 90 patients were treated with irradiation alone. In the chemotherapy/RT group, 32-month survival was 63% for stage IIB and 50% for III and in the RT group, 59% for IIB and 27% for IIIB tumors (differences not statistically significant). There was no difference in radiation-induced toxicity between the two groups.

In a Swedish study (587), 47 patients with carcinoma of the cervix were randomized to be treated with irradiation alone (64.8 Gy, 1.8-Gy fractions) and 47 with a combination of three cycles of cisplatin and 5 days of 5-FU administered every third week, followed by the same pelvic irradiation. The 5-year disease-free survival rates were 70% with chemoirradiation and 57% with irradiation alone (p = 0.07). The incidences of pelvic recurrence were 60% and 47%, respectively, and for distant metastasis, 19% and 35%, respectively. Two patients in the chemoirradiation and one in the irradiation-alone group died as a consequence of therapy.

Tseng et al. (620) published results of a study in which patients with advanced carcinoma of the cervix were randomly assigned to either RT alone (60) or concurrent chemotherapy (cisplatin, vincristine, and bleomycin every 3 weeks for a total of four courses) and RT (62). After a median follow-up of 46.8 months, the disease-free survival and actuarial survival rates were 51.7% and 61.7% in the concurrent group and 53.2% and 64.5% in the RT group, respectively (p = 0.27). Treatment-related toxicity was higher with the combination therapy compared with irradiation alone (36.7% vs. 17.7%; p = 0.02).

Randomized Studies in the United States

Results from several cooperative oncology groups demonstrated that cisplatin-based chemotherapy, when given concurrently with RT, prolongs survival in women with locally advanced cervical cancers, as well as in women with stage I to IIA disease who have metastatic disease in the pelvic lymph nodes, positive parametrial disease, or positive surgical margins at the time of primary surgery.

However, Curtin et al. (104) completed a phase III trial in which 89 patients with high-risk stage IB or IIA undergoing radical hysterectomy and pelvic node dissection were randomized to be treated with postoperative cisplatin/bleomycin alone (44 patients) or combined with pelvic RT (45 patients). There were nine and 10 recurrences, respectively, and survival was equivalent.

The details of five other published studies were summarized by Lehman and Thomas (370) (Table 66.28) and discussed in detail.

The GOG conducted randomized Protocol 85 in which patients with carcinoma of the cervix, a clinical stage of IIB to IVA, and negative para-aortic nodes were treated with external pelvic irradiation (51 Gy) combined with 30 Gy to point A with LDR brachytherapy (653); 177 patients received 5-FU (IV infusion, 1 g/m2 for 4 days) and cisplatin (50 mg/m2 IV) on days 1, 29, and 30 to 33, and 191 patients received hydroxyurea (80 mg/kg orally twice weekly). With a median follow-up for survivors of 8.7 years, the 5-year survival rate in the cisplatin/5-FU arm was 60%, compared with 47% for women in the hydroxyurea arm.

After completion of GOG 85, the group opened GOG 120 for the same patient population (654), which was a three-arm randomized trial comparing irradiation plus hydroxyurea versus irradiation plus weekly cisplatin versus irradiation plus hydroxyurea, cisplatin, and 5-FU. In 526 evaluable patients with a median follow-up for survivors of 35 months, the 4-year survival rate for women in both the weekly cisplatin and irradiation arm and the irradiation, 5-FU, cisplatin arm was 69%, compared with 37% the hydroxyurea and irradiation arm (p <0.001) Overall survival was also significantly better in the two patient groups receiving cisplatin. Hematologic toxicity was greater in the group treated with the three drugs compared with cisplatin or hydroxyurea alone.

The RTOG conducted a randomized study of 388 patients with stage IB to IIA larger than 5 cm, proven positive pelvic lymph nodes, or stage IIB to IVA carcinoma of the cervix in which patients were treated with either pelvic and para-aortic irradiation (best arm of RTOG Protocol 79-20) or pelvic irradiation and three cycles of concomitant chemotherapy with cisplatin (75 mg/m2) and 4-day infusion of 5-FU (1,000 mg/m2/ per day) (466).

Results were updated by Eifel et al. (145). With a median follow-up of 6.6 years for 228 survivors, the 8-year overall survival rate for women on the irradiation and cisplatin/5-FU arm was 67% versus 41% in the irradiation-only arm (p <0.0001; Fig. 66.34). Disease-free survival rates were 66% and 36%, respectively. There were no significant differences in late complications in the treatment groups.

Southwest Oncology Group 8797 was a study for women with FIGO stage IA2, IB, or IIA carcinoma of the cervix with metastatic disease in the pelvic lymph nodes, positive parametrial involvement, or positive surgical margins at the time of primary radical hysterectomy with total pelvic lymphadenectomy. Patients had confirmed negative para-aortic lymph nodes; if the para-aortic lymph nodes were not sampled, the patients had confirmed negative common iliac lymph nodes. One hundred twenty-seven patients were randomized to be treated with pelvic EBRT with 5-FU infusion and cisplatin, and 116 were

P.1594

treated with irradiation alone. The 3-year survival for women on the adjuvant cisplatin/5-FU and RT arm was 87% compared with 77% for women on the pelvic irradiation arm (Fig. 66.35) (483). The difference is statistically significant.

In GOG 123, 183 women with bulky (≥4 cm) stage IB carcinoma of the cervix with negative pelvic and para-aortic nodes radiographically or surgically determined were randomized to be treated with pelvic EBRT and brachytherapy, followed by extrafascial hysterectomy, and 186 received EBRT and brachytherapy with weekly cisplatin (40 mg/m2, total dose not to exceed 70 mg per week) followed by extrafascial hysterectomy (312).

With a median follow-up for survivors of 35.7 months, the progression-free survival rate for women treated with irradiation and cisplatin was 79% compared with 63% for those treated with RT alone (p <0.001). The overall 3-year survival rates were 83% and 74%, respectively (p = 0.0008).

The results of these five randomized trials were summarized by Rose and Eifel (524) (Table 66.29).

Haie-Meder et al. (221), in a review of these trials, concluded that the published results have led to a modification in the standard of treatment in these poor-prognosis cervix cancers. Five of the randomized trials evidenced the superiority of cisplatin-based chemotherapy, but the optimal chemotherapeutic regimens remain to be defined.

In GOG Protocol 165, patients with stages IIB to IVA cervical cancer received either radiation therapy and concurrent weekly cisplatin (40 mg/m2) or radiation therapy and a protracted venous infusion (PVI) of 5-FU.

Lanciano et al. (347) reported that the study was prematurely closed after an interim analysis showed a failure rate 35% higher and would not result in improved DFS with PVI 5-FU/RT compared with weekly cisplatin.

P.1595

Randomized Studies in Other Countries

Lorvidhaya et al. (387) reported on 673 patients with predominantly stage IIB and IIIB disease randomized to receive either irradiation alone or combined with chemotherapy administered in either an adjuvant, concurrent, or adjuvant and concurrent schedule. Concomitant chemotherapy consisted of mitomycin C (10 mg/m2) given on days 1 and 30 and oral 5-FU (300 mg/m2 per day) given on days 1 to 14 and 42 to 56. Adjuvant chemotherapy consisted of three cycles of oral 5-FU (200 mg per day) given for 4 weeks, with a 2-week rest every 6 weeks. With a median follow-up of 25 months, there was a statistically significant improvement in disease-free survival for all patients who received chemotherapy/RT, regardless of the timing of administration of the chemotherapy. However, the authors did not state the radiation dose delivered with brachytherapy, the total dose prescribed to point A, or the overall treatment time. In the absence of this information, the adequacy of the radiation therapy cannot be evaluated, and we cannot assume that the results of this study apply to all patients treated with irradiation.

On the other hand, Pearcey et al. (466) reported on a Canadian randomized study in which 127 patients with stage IB, IIA >5 cm, or IIB carcinoma of the cervix were randomized to be treated with cisplatin (40 mg/m2 weekly) and RT, and 126 patients were treated with RT alone (50.4 Gy to the pelvis combined with brachytherapy). With a median follow-up of 65 months, the 5-year survival rates were 59% and 56%, respectively (p = 0.43). There was a somewhat greater incidence of significant late morbidity in the RT alone group (12% vs. 6%; p = 0.08).

Possible explanations for the discrepancy in results between the five U.S. trials (436) and the NCI of Canada Study were analyzed by Lehman and Thomas (370).

Chemotherapy and Hyperfractionated Irradiation

Calkins et al. (66) assessed the toxicities of multiple–daily-fractionated (twice daily, 1.2-Gy fractions) whole pelvis radiation plus concurrent chemotherapy for locally advanced carcinoma of the cervix. In the first study (GOG 8801) of 38 patients, hydroxyurea was given orally (80 mg/kg to a maximum of 6 g) at least 2 hours before irradiation, twice every week. In the second

P.1596

study (GOG 8901) of 30 patients, cisplatin and 5-FU were used concomitantly with RT. Acute toxicity was primarily enteric and appeared to be dose related. The maximum tolerated dose of whole pelvis radiation that could be delivered in a hyperfractionated setting with concomitant chemotherapy was 57.6 Gy in 48 fractions followed by brachytherapy.

Thomas et al. (607) conducted a four-arm study in which 234 women with bulky stage IB to IVA cervical cancer were randomized to receive either standard RT (EBRT and brachytherapy to deliver 90 Gy to point A) with or without a 4-day infusion of 5-FU (1 g/m2) on days 1 to 5 and 22 to 25, or partially hyperfractionated RT with or without the same chemotherapy regimen. The partially hyperfractionated regimen delivered two fractions, 6 hours apart, on the first 4 days of treatment, coinciding with the infusion of 5-FU. The addition of 5-FU did not improve pelvic tumor control (37% to 75% at 5 years) or overall survival. However, this study closed without reaching target patient accrual.

Neoadjuvant Chemotherapy

Thomas (609) summarized the rationale and potential limitations of neoadjuvant chemotherapy in carcinoma of the cervix and pointed out that four randomized trials of neoadjuvant chemotherapy and irradiation had been reported. Although response rates to the chemotherapy are between 30% and 85%, none of the studies showed an advantage for pelvic tumor control or survival (33,70,77,244,341,361,587,593).

Colombo et al. (92), in a review of the literature, concluded that the role of neoadjuvant chemotherapy followed by radiation and by concomitant chemotherapy is controversial because high response rates are reported but no significant advantages in survival or local control have been shown.

Amifostine

De Los Santos and Small (110) reviewed the available data in several small phase I or II trials and concluded that this radioprotector may be beneficial in the treatment of patients with cervical cancer, particularly when using chemoradiation. Subcutaneous administration may facilitate the administration of the drug and decrease acute and late toxicity. RTOG conducted a phase I and II study (RTOG 0116) with subcutaneous amifostine in patients treated with extended field RT brachytherapy and chemotherapy without (Phase I) or with amifostine (Phase II). Phase I completed accrual with 27 patients and confirmed the significant morbidity of this combination therapy. The phase II results have not been reported yet.

Hyperbaric Oxygen and Hypoxic Sensitizers

Several reports evaluated the efficacy of hyperbaric oxygen combined with irradiation in the treatment of a variety of human tumors, including carcinoma of the uterine cervix.

Watson et al. (643), in a randomized clinical trial of 320 patients (stages III to IVA), reported a 5-year survival rate of 33% in the oxygen-treated group in contrast to 27% in the control group treated in normal air (p = 0.08). The local recurrence rate was 33% in the 161 patients treated with oxygen and 53% in 159 patients treated in normal air (p <0.001). Morbidity in the patients treated with oxygen was greater (20 severe and 13 moderate complications) than in those treated in normal air (six severe and eight moderate complications, respectively). The difference was particularly striking in the bowel (13 and two severe complications, respectively).

Dische et al. (124) reviewed the data in a randomized study of patients with advanced carcinoma of the cervix treated with radiation therapy and hyperbaric oxygen or air and noted that the patients treated with oxygen had improved survival at Mount Vernon and Glasgow but not at Cape Town. Data from the three centers were merged, and analysis showed that local tumor control was significantly worse in patients treated in normal air who had a prior blood transfusion, but in the oxygen group this effect was reversed. The same interaction was noted in the survival results (p = 0.042).

A trial reported by Fletcher (161), in which 233 patients with stage IIB, III, and IV carcinoma of the cervix were randomized to be treated with irradiation in normal air or with hyperbaric oxygen, demonstrated no significant benefit in survival or tumor control (20/109 patients treated with oxygen failed in the pelvis, in contrast to 29/124 treated in normal air). Furthermore, morbidity was greater (26 complications) in patients treated with hyperbaric oxygen compared with the control group (15 complications).

Dische et al. (124) published results of a four-arm randomized trial of hyperbaric oxygen and radiation therapy of stage IIB and III carcinoma of the cervix in which 335 patients were randomized to treatment in 10 or 28 fractions, in hyperbaric oxygen or in normal

P.1597

air. Data of 327 cases were analyzed. There was no advantage in tumor control with the use of hyperbaric oxygen. There was an increase in late radiation morbidity when treatment was given in hyperbaric oxygen rather than in normal air, and when using 10 fractions, a total dose of 45 Gy rather than 40 Gy was administered.

No definite conclusions can be drawn concerning the use of hyperbaric oxygen in carcinoma of the cervix. It is possible that hyperbaric oxygen administered with fewer high-dose fractions may be more efficacious than when combined with conventional dose and fractionation schemes. The trials reported have not shown an increased incidence of distant metastasis, which has been observed in a clinical study and in some animal experiments (286).

Grigsby et al. (210) published results of an RTOG study in which 120 patients with carcinoma of the cervix were randomized to receive irradiation alone or combined with misonidazole. The 5-year progression-free survival was 22% and 29%, respectively.

These findings are similar to those reported by Overgaard et al. (456), who, in a randomized study of 331 patients with carcinoma of the cervix treated with either misonidazole or a placebo and irradiation, found no significant difference in local tumor control (50% vs. 54%), disease-free survival (47% vs. 46%), or crude survival (39% vs. 45%).

Irradiation and Hyperthermia

Because of technical limitations in the delivery of adequate heat to large parts of the body such as the pelvis, the use of hyperthermia in the treatment of carcinoma of the uterine cervix has been sparse.

Hornback et al. (260) described a nonrandomized study in which the combination of microwave hyperthermia (433 MHz) and irradiation resulted in improved pelvic tumor control (72%) in a group of 79 patients with stage IIIB carcinoma compared with previously irradiated historic control patients (53%). However, 5-year survival rates were comparable in both groups (22% to 30%).

Sharma et al. (552) reported a 70% disease-free survival rate at 18 months in 20 patients with stage IIB or III carcinoma of the uterine cervix treated with a combination of irradiation and hyperthermia (13.5 MHz, 42°C to 43°C, 30 minutes before irradiation), in comparison with a 50% disease-free survival rate in 22 patients treated with irradiation alone. The grade 3 complication rate (8%) was similar in both groups.

Dinges et al. (123) treated 18 patients with advanced carcinoma of the cervix with RT plus hyperthermia (in the first and fourth weeks, two regional hyperthermia treatments were applied). The acute toxicity was low and similar to that with RT alone. The local tumor control was 48% at 2 years.

Harima et al. (232) evaluated irradiation therapy or thermoirradiation (three sessions of hyperthermia) for stage IIIB cervical carcinoma; two groups of 20 patients each were randomly divided. A complete response was achieved in 50% (10/20) in the RT group versus 80% (16/20) in the thermoirradiation group (p = 0.048). The 3-year overall survival and disease-free survival rates for the patients who were treated with thermoirradiation (58.2% and 63.6%) were better than with RT (48.1% and 45%), but differences were not statistically significant. The 3-year local relapse-free survival rate of the patients who were treated with thermoirradiation (79.7%) was significantly better than that of the patients treated with irradiation alone (48.5%; p = 0.048). Thermoirradiation was well tolerated and did not add to either acute or long-term toxicity over radiation alone.

Vasanthan et al. (632) reported on 110 patients with locally advanced cervix cancer randomized to treatment with RT alone or combined with hyperthermia (minimum five sessions, 60 minutes each, once weekly). Overall 3-year pelvic tumor control was 68.5% and survival 73.2%, with no difference in either group, although survival was lower in the patients with stage IIB treated with hyperthermia. Acute toxicity was 18% (10/55) in the hyperthermia patients and 4% (2/55) with RT alone. Late toxicity was no different in the two arms.

Carcinoma of the Cervix and Pregnancy

The concurrent presence of carcinoma in situ or invasive carcinoma of the uterine cervix and pregnancy, although rare, poses a therapeutic dilemma to gynecologic and radiation oncologists. Reported incidence is one to 10 cases per 10,000 pregnancies (128). In the United States, the incidence has decreased with the less frequent diagnosis of invasive cervical cancer.

Norstrom et al. (446), in Sweden, found cervical cancer was diagnosed in 33 women in association with pregnancy (incidence 11.1 cases per 100,000 deliveries and 7.5 per 100,000 pregnancies). Abnormal bleeding was the symptom that led to diagnosis in 54.5% of the women; 45.5% were asymptomatic but had an abnormal cervicovaginal cytologic test result (39.4%) or abnormal vaginal examination (6.1%) in association with pregnancy. During the follow-up, 1/12 women with cervix cancer in the first trimester, 4/12 in the third trimester, and 2/9 women postpartum died of the disease. Primary surgery was used more frequently than radiation therapy.

For carcinoma in situ, if the pregnancy is allowed to reach full term, confirmation of the diagnosis by colposcopy and conservative management with monthly Pap smears constitutes the best approach. Conization has frequently been performed. Punch biopsies can be obtained, but the diagnostic accuracy is less reliable. As many as 50% of the patients have residual carcinoma in situ after delivery.

In patients with invasive carcinoma, the lesion is usually clinically apparent. Multiple punch biopsies are adequate to confirm the diagnosis.

Management is individualized based on tumor size and stage, patient age, and desires of the patient (or couple) regarding the pregnancy. The majority of patients with cervical cancer diagnosed during pregnancy (approximately 75%) have stage I tumors (128,258,416).

Occasionally in late pregnancy, if tumors are small, definitive therapy is postponed and delivery is allowed. In a review of the literature intentional treatment delay was associated with a recurrence rate of 4% (505).

Sorosky et al. (572) reported on eight pregnant women with stage I carcinoma of the cervix who had declined immediate therapy and followed until the late third trimester; a cesarean section–radical hysterectomy was performed with delay in therapy ranging from 3 to 40 weeks during the pregnancies. There was no clinical progression of the disease with follow-up of 33 months.

Greer et al. (192) noted in 600 infants without congenital abnormalities, when stage IB cervical carcinoma was diagnosed during pregnancy and fetal survival was desired, the neonatal mortality rate decreased from 30% when the fetus was delivered at 26 to 27 weeks to 2.7% when the fetus was allowed to mature to 34 to 35 weeks.

Because there is a greater need to institute therapy as soon as possible that the accepted method of treatment in patients in the first 6 months of pregnancy is to carry out definitive surgery or radiation therapy, as indicated by the stage of the disease. In the third trimester of pregnancy, when the fetus may be salvaged, some gynecologic oncologists prefer a postpartum cesarean section, combined with a radical hysterectomy and lymphadenectomy or followed by definitive treatment. However, some authors report that vaginal delivery has no detrimental effect on the prognosis (100).

If a radical hysterectomy is performed and positive pelvic lymph nodes are found, the usual postoperative irradiation, including external beam with or without intracavitary insertion, should be carried out.

If it is decided to terminate the pregnancy and treat the patient with external irradiation, initially the whole pelvis is irradiated (40 Gy in 4 weeks). A spontaneous abortion occurs, and there is some involution of the uterus. However, in one

P.1598

series of 45 patients 27% did not abort spontaneously and surgical evacuation was required (568). After this dose of radiation, careful evacuation of the uterus and LDR (or equivalent dose HDR) brachytherapy may be performed under general anesthesia. Two intracavitary insertions for a total of 6,500 mgh (55 Gy to point A) and an additional 10 or 20 Gy are delivered to the parametria with a midline block. If surgery is to be carried out, approximately 4,000 mgh is given.

Patients who require high doses of pelvic irradiation (>40 Gy) should be counseled regarding the permanent loss of reproductive capability, not only because of ovarian ablation (which happens with doses of 8 Gy or higher), but as a consequence of radiation effects in the uterus (570).

Survival is the same regardless of the trimester of the pregnancy in which definitive treatment is instituted. Creasman et al. (100) reported on 48 patients treated by irradiation, 45 by irradiation followed by surgery, and five with radical hysterectomy alone. The survival was comparable with that of nonpregnant patients for similar stages. The survival rate for patients with stage I disease was comparable whether vaginal delivery was allowed or a cesarean section was performed (approximately 85% in stage I and 50% to 64% in stage II). Also, the percentage of infants surviving (>80%) was the same in both groups.

Sood et al. (570) performed a retrospective analysis of 26 women with cervical carcinoma diagnosed during pregnancy and treated primarily with radiation therapy (mean dose, 46.7 Gy) and LDR intracavitary radiation (mean dose, 56.5 Gy to point A). These cases were matched with 26 nonpregnant control patients based on age, histology, stage, treatment, and year of treatment. There were no statistically significant differences in recurrence rates or survival between the pregnant group and the control patients. Short-term toxicity was comparable in pregnant and nonpregnant patients. Long-term complication rates were 12% in pregnant patients and 27% in control patients, but this difference was not statistically significant. Most complications were likely related to radiation techniques (particularly the predominance of 60Co).

Sood et al. (569) compared the prognosis of 56 women who had cervical cancer diagnosed during pregnancy and 27 who were diagnosed within 6 months after delivery. Control patients (cervical cancer diagnosed at least 5 years since last delivery) were matched one-to-one with cases based on age, histology, stage, treatment, and time of treatment. Among the postpartum women, 11 were treated with radical hysterectomies, 14 with radiation therapy, and two with stage IA1 disease were treated with vaginal hysterectomies. One of seven patients who had cesarean sections had a local and distant recurrence. In contrast, 10/17 (59%) who delivered vaginally had recurrences (p = 0.04). In multivariate analysis, vaginal delivery was the most significant predictor of recurrence, followed by high tumor stage. Survival for patients diagnosed in the postpartum period was significantly worse than for control patients and for those diagnosed during pregnancy. They concluded that pregnant women with cervical cancer should be delivered by cesarean section.

Jones et al. (288) published a survey by the American College of Surgeons that evaluated management of invasive cervical carcinoma in 161 pregnant patients; 86 were treated with surgery alone, 30 with radiation therapy alone, and 45 with a combination of the two modalities. Approximately one-third of patients were diagnosed in each trimester. The 5-year survival was 94.6% for patients diagnosed in the first trimester, 76.9% for the second, and 68.9% for the third trimester. The prognosis of patients with invasive carcinoma of the cervix associated with pregnancy was similar to that of nonpregnant patients. There was no significant difference in 5-year survival whether the patients delivered by cesarean section or normal vaginal delivery.

Senekjian et al. (549) reported no difference in survival or patterns of failure in 24 women who were pregnant at the time of diagnosis of clear-cell adenocarcinoma of the cervix and vagina compared with 408 who had never been pregnant. The 5- and 10-year actuarial survival rates were 86% and 68% for the pregnant patients and 87% and 79%, respectively, for patients who had not been pregnant.

Neoadjuvant chemotherapy for invasive squamous-cell carcinoma of the cervix in pregnancy has been considered in selected patients (604).

The practice popularized 30 years ago of administering a “restraining dose of radium” and deferring definitive radiation therapy until delivery is carried out should be strongly condemned. Strauss (583) reported 2/11 infants being born with microcephaly in addition to other complications such as alopecia, facial deformity, eye damage, and chromosomal abnormalities after this procedure.

Carcinoma of the Cervical Stump

Subtotal hysterectomy, a relatively popular procedure for benign conditions of the uterus in past years, is rarely performed today. These patients are, of course, at risk for development of carcinoma of the uterine cervix.

It is important to divide carcinoma of the cervical stump into true, when the first symptom occurs 3 or more years after subtotal hysterectomy, or coincidental, when the symptoms are noticed before the third postoperative year. This separation is important because the prognosis for carcinoma of the true stump is significantly better than for coincidental lesions, in which carcinoma was probably present when the hysterectomy was performed (478).

The natural history and patterns of spread of carcinoma of the cervical stump are similar to those of the cervix in the intact uterus. The diagnostic work-up, clinical staging, and basic principles of therapy are the same.

When surgery is performed for stage I tumors, it is somewhat more difficult because of the previous surgical procedures and the presence of adhesions in the pelvis. When irradiation is administered, the lack of uterine cavity into which to insert a tandem containing two or three sources makes intracavitary therapy more difficult. As many sources as technically feasible should be inserted in the remaining cervical canal. Occasionally, transvaginal irradiation may be used to boost the dose delivered to central disease in the stump. It is important to deliver higher whole pelvis irradiation. In general, patients with stage I disease are treated with a combination of 20 Gy to the whole pelvis and 30 Gy to the parametria with midline shielding combined with two intracavitary insertions. The dose of intracavitary therapy depends on the number of sources that can be placed in the cervical canal. More advanced stages should be treated with 40 Gy to the whole pelvis and 20 Gy to the parametria with midline shielding, combined with the same intracavitary doses. When there is no opportunity to insert any sources in the cervical canal, the whole pelvis dose must be increased to 60 Gy. Total dose (external and LDR intracavitary brachytherapy) to the upper vaginal mucosa should not exceed 150 Gy, and tolerance doses to small volumes of the bladder (80 Gy) or rectum (75 Gy) should be carefully monitored.

If there is bulky disease present in the cervix, parametrium, or vagina, additional interstitial therapy is advisable, if technically feasible. When intravaginal cones are used, a 30- to 40-Gy air dose is delivered in 2 to 3 weeks, in three to five weekly fractions.

The 5-year survival rate for carcinoma of the cervical stump treated with irradiation is similar to that reported for patients with carcinoma of the intact uterus (336,657). The anatomic sites of failure and the incidence of recurrences are similar to

P.1599

those of patients in whom the uterus is intact. Distant metastases also follow the same distribution.

In 253 patients with carcinoma of the cervical stump treated at M.D. Anderson Cancer Center, median survival was 203, 140, and 32 months for stages I, II, and III, respectively (411).

Kovalic et al. (336) reported on 70 patients with carcinoma of the surgical stump treated with irradiation; 16 also underwent a surgical procedure. The 10-year disease-free survival was 79% for stage IB, 66% for stage IIB, and 39% for stage IIIB disease. The pelvic failure rates were 10%, 9%, and 50%, respectively. Major gastrointestinal complications were noted in 9% of patients, and urinary complications in 3.8%. The results are comparable with those seen in patients treated for invasive carcinoma of the cervix with intact uterus.

Hannoun-Levi et al. (229) published results in 77 patients treated for carcinoma of the cervical stump. Treatment consisted of a combination of EBRT and brachytherapy, and, in a few cases, patients underwent surgery or interstitial brachytherapy. Three-year pelvic tumor control was achieved in 59/77 patients (76.6%); tumor control probabilities were 77%, 73.7%, and 56% in patients with stage I, II, or III tumors, respectively. Late complications were grade 2 in five patients (6.5%); grade 3 in one patient (1.3%), and grade 4 in two patients (2.6%).

Hellstrom et al. (240) published a retrospective study of 145 patients treated for carcinoma of the cervical stump, representing 2.2% of all cervical cancers. Three control cases to each case were matched from the cohort of cases with cervical carcinoma with intact uterus. The dose of irradiation from the intracavitary application given to the stump cancers was lower than for comparable cases with intact uterus. Long-term prognosis for squamous-cell carcinoma of the uterine stump was comparable with that of the ordinary cervical carcinomas. Stump adenocarcinomas had a worse prognosis compared with adenocarcinoma of the intact uterus (p <0.07) and with stump of the squamous-cell carcinoma (p = 0.05). The complication rate was higher for stump cancer cases compared with that for cervical cancers with intact uterus.

Because of the close proximity of the bladder, rectum, and small intestine to the intracavitary sources and owing to the often higher doses of external-beam irradiation given to the whole pelvis, complications are somewhat more frequent than in carcinoma of the cervix with intact uterus.

The use of concurrent chemotherapy with external-beam irradiation in patients with cervical stump carcinoma may be beneficial, especially because, as discussed, the brachytherapy dose is often compromised by anatomic constraints.

Second Malignancy

The possible induction of secondary primary cancers by pelvic irradiation is controversial (616).

Lee et al. (367) observed no significant increase in the incidence of second malignancies in patients irradiated for carcinoma of the cervix in comparison with the Connecticut Tumor Registry prevailing rates.

Boice et al. (43), in a review of 68,730 women with carcinoma of the cervix treated with radiation therapy, observed a second malignant tumor in 3,324, compared with 3,063 expected (4.8% increase; p <0.001). The excess was concentrated in the lung, other genital organs, bladder, and rectum. In addition, in 10,817 women with invasive cervical cancer not treated with irradiation, 479 secondary malignant tumors were observed versus 435 expected (4.4%; p = 0.02). Thus, the incidence of secondary tumors in women treated for carcinoma of the cervix with or without irradiation is only slightly greater than in the general population. Pelvic organs receiving a high dose of irradiation appear to have a somewhat greater incidence of a second primary.

Storm (581), in a comprehensive analysis of the Danish Cancer Registry data of 24,970 women with invasive cervical cancer and 19,470 with carcinoma in situ of the cervix treated between 1943 and 1982, noted a small overall excess of secondary primary cancers in the lung, stomach, pancreas, rectum, and bladder and connective tissue sarcomas, although there was a decreased incidence of breast cancer in the irradiated patients compared with nonirradiated patients (attributable to ovarian ablation by radiation therapy). In the patients irradiated for invasive carcinoma, there was an excess of 64 cases per 10,000 women per year of tumors in organs close to or at an intermediate distance from the cervix, reaching a maximum after 30 years or longer of follow-up. A high risk for development of acute nonlymphatic leukemia was observed in irradiated patients with carcinoma in situ, but not in those with invasive lesions. This could be explained by the lower doses of irradiation delivered to the bone marrow in the in situ tumors treated with brachytherapy alone, with greater induction of mutations and less cell killing, which may be responsible for the leukemogenic effect. Decreased risk was noted for tumors of the brain, myeloma of the skin, and tumors of the colon other than rectal.

In a study of 117,830 women diagnosed with cervical carcinoma in situ and 17,556 with invasive cervical carcinoma in Sweden, treatment not specified and in situ lesions traditionally treated with surgery alone, there was an increased incidence (RR 2.3 to 3) of second primary tumors in the anus, rectum, urinary bladder, pancreas, esophagus, and lung compared with the standardized incidence rate for all women (241). The data showed consistent increases in suggested targets for HPV at tobacco-related sites. A contributing role for a depressed immune response was considered to play a role.

Werner-Wasik et al. (649), in an analysis of 125 women with stage I and II carcinoma of the cervix treated with radiation therapy, observed 11 secondary primary tumors in 10 patients (four breast, two lung, and one each of myeloma, non-Hodgkin's lymphoma, bladder, thyroid, and vulva). All secondary primary tumors were located outside the irradiation fields.

The increased relative risk of breast cancer in these patients was 2.64, higher than reported by Boice et al. (43).

Mark et al. (401) identified 13/114 patients diagnosed with uterine sarcoma who had a prior history of pelvic irradiation (doses of 40 to 80 Gy). Criteria for radiation-induced sarcomas included

• Previous history of pelvic irradiation,

• Latent period of several years,

• Development of sarcoma within previously irradiated field, and

• Histologic confirmation of malignancy.

Histologic types of tumor were mixed müllerian in six, leiomyosarcoma in four, endometrial stroma sarcoma in one, fibrosarcoma in one, and angiosarcoma in one patient. Sarcoma developed in the uterus in 12 and at the vaginal cuff in one. Ten patients were treated with surgery and two with radiation therapy. The 5-year disease-free survival rate after salvage therapy was 17%.

Cost of Care for Cervical Cancer

Wolstenholme and Whynes (660) carried out a detailed cost audit over 5 years on a sample of patients diagnosed in 1990 in one United Kingdom region. The mean costs of managing preinvasive carcinoma were found to be significantly lower than those of stage I invasive carcinoma, and both were lower than the costs of invasive cancer at stages II to IV.