Management of recurrent or metastatic cervical cancer

INTRODUCTION — Although uncommon at initial diagnosis, metastatic disease will develop in 15 to 61 percent women with cervical cancer, usually within the first two years of completing treatment [1,2]. In the majority of cases, metastatic cervical cancer is not curable. However, for some patients who present with recurrent disease in the pelvis (locoregional recurrence) or with limited distant metastatic disease, surgical treatment is potentially curative.

Histologically, squamous cell carcinoma (SCC), adenocarcinoma, and adenosquamous carcinomas account for approximately 70, 25, and 3 to 5 percent of all cervical cancers, respectively (table 1) [2]. Rarer histologies include neuroendocrine or small cell carcinomas.

The management of recurrent or metastatic cervical cancer will be discussed here. Specific issues regarding patterns of relapse after treatment are discussed separately. (See "Invasive cervical cancer: Patterns of recurrence and post-treatment surveillance".)

CLINICAL PRESENTATION — Recurrent cervical cancer can present as a local recurrence or as metastatic disease.

Locally recurrent cervical cancer usually presents with vaginal symptoms (ie, discharge, bleeding, dyspareunia, or pain). On pelvic exam, a mass or nodularity at the vaginal cuff, which may extend to the pelvic side wall, may be visualized or palpated. Disease within the vagina (or vaginal vault) can be tender to palpation and/or prone to bleed easily.

By comparison, patients with metastatic cervical cancer usually present with either no symptoms or nonspecific complaints (ie, fatigue, nausea, or weight loss); however, they may have symptoms related to the site of metastases. As an example, patients with bone metastases present with chronic pain, not alleviated with rest.

DIAGNOSTIC EVALUATION — Patients with findings suspicious for recurrence on pelvic exam should undergo a biopsy.

Women who present with signs (ie, weight loss, palpable abdominal lesions) or symptoms (ie, pain, nausea, or weight loss) should undergo radiologic imaging to evaluate for metastatic disease. In addition, imaging is essential for women who present with a mass on pelvic exam because the extent of disease will influence the treatment plan.

Metastatic cervical cancer may present as nodal disease involving the pelvic, para-aortic, and/or supraclavicular nodes; limited disease involving one organ site; or widely metastatic disease. In one study of over 200 patients, distant metastases, disease limited to the para-aortic nodes, and disease in the supraclavicular nodes (with or without para-aortic node involvement) was found in 65, 22, and 14 percent, respectively [3]. In another review of the literature, the estimated prevalence of organ site involvement by metastatic cervical cancer was [4]:

●Pelvic or para-aortic nodes (75 and 62 percent, respectively)

●Lung (33 to 38 percent)

●Liver (33 percent)

●Peritoneum (5 to 27 percent)

●Adrenal gland (14 to 16 percent)

●Intestines (12 percent)

●Skin (10 percent)

The most commonly used imaging modalities to evaluate recurrent cervical cancer include computed tomography (CT) and positron emission tomography (PET) with or without CT. Of these, we suggest PET/CT for evaluation of local and distant disease. It has a sensitivity of 93 to 96 percent and specificity of 93 to 95 percent [5]. In addition, the results from a PET/CT scan often lead to changes to the therapeutic plan for women with recurrent disease by sparing women from an extensive surgical procedure in the setting of widely metastatic disease [6-8].

Patients presenting with isolated metastatic findings on imaging should undergo a biopsy to prove metastatic disease, as there is a risk that these findings may represent a second primary malignancy or a benign process. As an example, in one series of 23 women with cervical cancer who presented with pulmonary lesions and underwent resection, seven (29 percent) were diagnosed with primary lung cancer at final pathologic analysis [9]. (See 'Limited metastatic disease' below.)

MANAGEMENT OF LOCAL RECURRENCE — For women who present with a local relapse confined to the cervix or vagina, treatment directed to the site of recurrence can be performed with curative intent. Options include hysterectomy or pelvic exenteration in patients who have received prior radiation therapy (RT), or RT in those who have not received RT or are not surgical candidates; the choice depends on the patient's prior treatment.

Patients who have previously been treated with RT and who are not candidates for surgical resection should be offered chemotherapy. The approach to these patients is identical to the treatment of women with metastatic disease. (See 'Metastatic disease' below.)

Candidates for surgical resection — Patients who experience a local recurrence should be offered surgical resection with curative intent. Commonly employed criteria to identify those women most likely to benefit from surgery include [10] (see "Invasive cervical cancer: Patterns of recurrence and post-treatment surveillance", section on 'Patterns of recurrence'):

●A central pelvic recurrence without side wall fixation or associated hydronephrosis

●A long disease-free interval

●Tumor size of the recurrence less than 3 cm in diameter

Surgical procedure

Following RT with uterine conservation — For women who underwent primary radiation therapy (RT), radical hysterectomy or pelvic exenteration for management of local recurrence is the preferred approach and is associated with five-year survival rates ranging between 30 and 40 percent [11,12]. However, surgical complications are more common in this setting. In one study, 15 of 34 patients who underwent surgery for persistent or recurrent disease following RT experienced major postoperative complications, including fistula formation (in five women).

Following hysterectomy — For women who have undergone hysterectomy (with or without adjuvant RT or chemoradiation), pelvic exenteration represents a potentially curative option for local recurrence or persistent disease. Careful patient selection is required given the perioperative and postoperative morbidity associated with this extensive surgical approach. Vaginectomy, rather than pelvic exenteration, has also been described in select patients [13]. (See "Exenteration for gynecologic cancer" and "Vaginectomy".)

Some patients treated with surgical resection for cervical cancer may benefit from intraoperative radiation therapy (IORT) where it is available. However, the efficacy of IORT has not been shown in prospective trials [14,15].

Nonsurgical options

Candidates for RT — Radiation therapy (RT) is a reasonable option for the following patients:

●Women who have not been previously treated with RT

●Women with operable disease who opt not to proceed with pelvic exenteration

The benefit of RT was demonstrated in a single-institution experience of 35 women who were treated with high-dose RT following a pelvic recurrence [16]. The 5- and 10-year survival rates were 43 and 33 percent, respectively, and pelvic control rates were 69 and 62 percent, respectively. The use of brachytherapy and a long treatment-free interval between primary surgery and diagnosis of recurrence were positive predictors of a good outcome.

Given the superiority of concomitant chemotherapy with RT (chemoradiation) over RT alone as primary treatment, most experts prefer chemoradiation for these patients. (See "Management of locally advanced cervical cancer", section on 'Primary chemoradiation'.)

Candidates for chemotherapy — For patients who are not candidates for surgical resection or salvage RT, chemotherapy is the only option. Chemotherapy has activity for the treatment of cervical cancer, although treatment is less successful if the recurrence is in an area that was previously irradiated. (See 'Initial treatment' below.)

DISEASE ISOLATED TO THE LYMPH NODES — For women presenting with disease limited to the nodes (para-aortic and/or supraclavicular nodes), the prognosis is poor. This was illustrated in a retrospective study of 375 patients with recurrent cervical cancer [3]. The rate of overall survival (OS) at five years was 27 and 0 percent for women with limited metastatic disease involving the para-aortic nodes (n = 60) or the supraclavicular nodes (n = 26) [3].

There are limited data to help guide treatment of women with metastatic disease limited to the lymph nodes. Some experts prefer systemic chemotherapy, while others prefer RT (with or without chemotherapy). The choice between treatments depends on institutional practice and patient preference. (See 'Initial treatment' below.)

LIMITED METASTATIC DISEASE — For patients with limited metastatic disease (ie, disease in the para-aortic nodes or solitary lung metastasis), treatment can be directed towards the area of recurrence using radiation therapy (RT) or, in selected cases, a surgical approach.

In selected patients who present with limited metastatic disease, resection may be feasible. As an example, in one series of 15 women who underwent resection of pulmonary metastases for cervical cancer, 13 were alive with a median follow-up of 18 months (range, 4 to 68 months) [9]. The short interval of follow-up and lack of historical controls (that received chemotherapy, for example) limits interpretation of this approach. Given the potential morbidity associated with resection, careful selection of patients is required. (See "Surgical resection of pulmonary metastases: Outcomes by histology", section on 'Gynecologic cancers' and "Surgical resection of pulmonary metastases: Benefits, indications, preoperative evaluation, and techniques".)

METASTATIC DISEASE — For the majority of patients with metastatic disease and those who are not candidates for a local treatment approach, we administer chemotherapy. For women being treated in the first-line setting, we administer chemotherapy plus bevacizumab rather than chemotherapy alone. Treatment incorporating bevacizumab was shown to improve OS in these patients. (See 'Initial treatment' below.)

There are several agents with activity in cervical cancer, which can be used as part of a combination regimen or as single-agent therapy.

Treatment for metastatic cervical cancer can be stratified on whether or not prior chemotherapy was administered. This stratification is based on a retrospective evaluation by the Gynecologic Oncology Group (GOG) that included 428 patients treated on first-line chemotherapy trials [17]. The following individuals had a higher likelihood of a poor response to cisplatin-based combination chemotherapy:

●Black people (versus non-Black people) – odds ratio (OR) for poor response 0.49 (95% CI 0.28-0.83)

●Those with performance status 1 or 2 (versus 0) – OR 0.60 (95% CI 0.38-0.94)

●Those with pelvic disease (versus nonpelvic location) – OR 0.58 (95% CI 0.38-0.90)

●Those with prior treatment with cisplatin as part of chemoradiation (versus none) – OR 0.52 (95% CI 0.32-0.85)

●Those with recurrence within one year of diagnosis (versus longer) – OR 0.61 (95% CI 0.39-0.95)

These data were used to construct a prognostic model to predict treatment response to cisplatin-based chemotherapy by identifying three risk groups: high-risk (four to five factors present), mid-risk (two to three factors present), and low-risk (zero to one factor present). This model was externally validated in a separate group of patients treated on a separate GOG trial with the following results [17]:

●High-risk group – The overall response rate (ORR) was 14 percent, median progression-free survival (PFS) was 3 months, and median OS was 5.6 months

●Mid-risk group – ORR 29 percent, PFS 4 months, OS 7.6 months

●Low-risk group – ORR 43 percent, PFS 7 months, OS 12 months

This model suggests that chemotherapy-naive patients have a higher response rate than women who received prior chemotherapy, including as part of chemoradiation [18,19].

For patients with metastatic, advanced, or recurrent cervical cancer who are candidates for chemotherapy, cisplatin is the most widely used agent [20]. However, there are no randomized trials comparing single-agent cisplatin with other single agents (eg, paclitaxel, gemcitabine, or vinorelbine). The approach to chemotherapy in the first- and later-line treatment setting is reviewed below.

Initial treatment

Components of treatment — For women with recurrent, metastatic, or advanced cervical cancer, we recommend treatment consisting of a platinum-based combination plus the angiogenesis inhibitor, bevacizumab, and an immune checkpoint inhibitor. However, the costs of therapy may require scrutiny in comparison to the benefits and risks of incorporating bevacizumab or immune checkpoint inhibitors in this setting, especially in low-resource settings. (See 'Addition of a checkpoint inhibitor' below.)

Combination chemotherapy, as backbone — In the first-line setting we suggest a platinum-based combination. In addition, the data support the use of a platinum-based combination rather than single-agent cisplatin. Our preferred combination is cisplatin plus paclitaxel rather than other cisplatin-based regimens. This recommendation is based on the data below (see 'Combination versus single-agent therapy' below):

●The GOG 240 trial randomly assigned 452 women with metastatic cervical cancer to cisplatin plus paclitaxel or to paclitaxel plus topotecan with a second randomization to treatment with or without bevacizumab [21]. Treatment with paclitaxel plus topotecan (with or without bevacizumab) resulted in:

•A significantly higher risk of progression compared with cisplatin plus paclitaxel (median PFS, 5.7 versus 7.6 months, respectively; HR 1.39, 95% CI 1.09-1.77)

•No significant impact on OS (median, 12.5 versus 15 months; HR 1.20, 99% CI 0.82-1.76)

The administration of bevacizumab for these women is discussed above. (See 'Initial treatment' above.)

●The GOG 204 enrolled 434 patients with metastatic or advanced cervical cancer and randomly assigned them to treatment with cisplatin plus paclitaxel (the reference control arm) or one of three experimental regimens (cisplatin plus vinorelbine, cisplatin plus gemcitabine, or cisplatin plus topotecan) with the following results [22]:

•There was no difference in response rates – For cisplatin plus paclitaxel, the ORR was 29 percent. The ORR was 26, 22, and 23 percent for cisplatin administered with vinorelbine, gemcitabine, or topotecan, respectively.

•There was no difference in the risk of death among any of the experimental regimens compared with cisplatin plus paclitaxel.

•Cisplatin plus gemcitabine resulted in much less serious febrile neutropenia compared with cisplatin in combination with paclitaxel, vinorelbine, or topotecan (6 versus 13, 14, and 10 percent), respectively. However, cisplatin plus paclitaxel resulted in less serious thrombocytopenia compared with the other arms (7 versus 8, 28, and 35 percent for cisplatin plus vinorelbine, gemcitabine, or topotecan, respectively).

Alternate use of carboplatin — Because of the toxicity seen with cisplatin-based combination chemotherapy, carboplatin is a reasonable substitute for cisplatin, particularly for patients with medical comorbidities (eg, pre-existing renal failure) and those patients previously treated with cisplatin-based chemoradiation. This is based on the results of the Japanese Clinical Oncology Group 0505 trial (JCOG 0505) [23]. In JCOG 0505, 253 women with stage IVB, persistent or recurrent cervical cancer were randomly assigned treatment with cisplatin (50 mg/m²) plus paclitaxel (135 mg/m²) or carboplatin (area under curve [AUC] 5) plus paclitaxel (175 mg/m²), administered every three weeks for six cycles. Prior cisplatin therapy (primarily with chemoradiation) was noted in 43 and 50 percent of each group, respectively. Compared with cisplatin plus paclitaxel, treatment with carboplatin and paclitaxel resulted in:

●Similar ORR (63 versus 60 percent).

●No difference in OS (HR for mortality 0.99, 90% CI 0.79-1.25).

●Significantly less serious (grade 4) neutropenic events (45 versus 75 percent, p <0.0001). There were also less serious (grade 3/4) incidences of renal insufficiency (0 versus 2.4 percent), nausea, and vomiting (3 versus 7 percent). However, carboplatin plus paclitaxel resulted in more thrombocytopenia (25 versus 3 percent) and sensory neuropathy (5 versus 0 percent).

A post-hoc analysis showed that prior platinum exposure may impact outcomes such that those women not previously treated with cisplatin had a lower OS when paclitaxel was combined with carboplatin rather than cisplatin (median, 13 versus 23 months; HR 0.69, 95% CI 0.47-1.02). There was no statistically significant difference among women who were previously treated with cisplatin (HR 0.69, 95% CI 0.47-1.02).

The results of JCOG 0505 establish carboplatin and paclitaxel as a reasonable alternative to cisplatin plus paclitaxel in the treatment of women with metastatic cervical cancer, particularly in those who are not candidates for cisplatin and/or were previously treated with cisplatin-based chemoradiation. A 2014 systematic review including 1181 patients similarly found that carboplatin was a reasonable alternative to cisplatin [24].

Combination versus single-agent therapy — The evidence to support the use of a platinum-based combination rather than single-agent therapy is limited. However, combination therapy was compared against single-agent cisplatin in a 2012 meta-analysis that included five randomized trials (n = 1114) [25]. Compared with combination platinum-based therapy, single-agent cisplatin resulted in a lower response rate (relative risk [RR] 0.60, 95% CI 0.44-0.81) but was associated with less toxicity, including less serious (grade 3/4) neutropenia (RR 0.04, 95% CI 0.02-0.12), and thrombocytopenia (RR 0.16, 95% CI 0.05-0.48).

In two trials, the combination of cisplatin plus paclitaxel resulted in OS ranging from 13 to 15 months and PFS from 6 to 8 months. In contrast, OS ranged from 7 to 9 months and PFS was approximately three months with cisplatin alone (>11 trials).

Addition of bevacizumab — The incorporation of bevacizumab is based on the results of GOG 240, in which 452 women with metastatic, persistent, or recurrent cervical carcinoma were randomly assigned to chemotherapy with or without bevacizumab. Previous platinum-based therapy was administered with RT in 75 and 74 percent of patients, respectively. Final results (the prespecified cutoff for final analysis [348 deaths] occurred) were nearly identical to the results reported after a 2012 interim analysis [21,26]. Chemotherapy plus bevacizumab resulted in:

●A significant improvement in OS compared with chemotherapy alone (median, 16.8 versus 13.3 months, respectively; hazard ratio [HR] 0.77, 98% CI 0.62-0.95).

•OS in women who did not receive prior RT was higher in both groups (24.5 versus 16.8 months).

•Post-progression survival was not significantly different across groups (8.4 versus 7.1 months; HR 0.83, 95% CI 0.66-1.05).

●A significant improvement in PFS (median, 8.2 versus 6.0 months; HR 0.68, 95% CI 0.54-0.82).

●A significant improvement in the ORR (49 versus 36 percent).

Treatment with bevacizumab resulted in a significantly increased rate of grade 2 or greater hypertension, defined as systolic 160 to 179 mmHg and/or diastolic 100 to 109 mmHg (25 versus 2 percent; risk ratio [RR] 13.8, 95% CI 5.1-37). Other serious (grade 3/4) toxicities included: gastrointestinal fistula (3 versus 0 percent); genitourinary fistula (3 versus <1 percent); neutropenia (36 versus 26 percent; RR 1.4, 95% CI 1.0-1.8); and thrombosis or embolism (8 versus 2 percent; RR 4.5, 95% CI 1.6-13). There was no difference in quality of life reported at up to nine months following the start of therapy [27].

Taken together, these results support the use of chemotherapy plus bevacizumab as a first-line treatment of metastatic cervical cancer, and in August 2014, bevacizumab received approval for this indication by the US Food and Drug Administration and the United Kingdom Cancer Drug Fund. The choice of the chemotherapy partner is discussed above. (See 'Combination chemotherapy, as backbone' above.)

Addition of a checkpoint inhibitor — For patients with metastatic cervical cancer or recurrent cervical cancer not amenable to curative localized therapies, we recommend addition of either atezolizumab (irrespective of programmed cell death ligand 1 [PD-L1] status) or pembrolizumab (for PD-L1 positive disease only). The choice between atezolizumab and pembrolizumab in PD-L1 positive disease should be made based on cost and availability, given a lack of direct comparisons of these two agents in this setting.

Atezolizumab — The addition of atezolizumab to platinum-based chemotherapy and bevacizumab has shown OS improvements in this setting, regardless of PD-L1 status. Pembrolizumab is another option for PD-L1 positive disease only and is discussed below. (See 'Pembrolizumab, as another option for PD-L1 positive disease' below.)

In an open-label randomized trial in 410 patients with metastatic or recurrent cervical cancer not amenable to curative treatment, the addition of atezolizumab to platinum-paclitaxel chemotherapy and bevacizumab improved both median PFS (13.7 versus 10.4 months; HR 0.62, 95% CI 0.49-0.78) and OS (32 versus 23 months; HR 0.68, 95% CI 0.52-0.88) at a median follow-up of 33 months [28]. Grade ≥3 events occurred in 79 percent receiving atezolizumab and 75 percent in the control group.

Although this approach does not yet have regulatory approval, we consider it to be an appropriate strategy based on these published results.

Pembrolizumab, as another option for PD-L1 positive disease — For patients whose cancers have PD-L1 expression combined positive score (CPS) ≥1, the addition of pembrolizumab to chemotherapy and bevacizumab has shown OS improvements and has regulatory approval in the US [29]. Either pembrolizumab or atezolizumab may be used in patients with PD-L1 positive disease. Atezolizumab is discussed above. (See 'Atezolizumab' above.)

In a phase III trial in 617 patients receiving first-line chemotherapy for persistent, recurrent, or metastatic cervical cancer (approximately two-thirds of whom also received bevacizumab), the addition of pembrolizumab improved median PFS versus placebo (10.4 versus 8.2 months; HR for disease progression or death 0.65, 95% CI 0.53-0.79), without negatively impacting health-related quality of life [30,31]. Objective response rates were 66 and 51 percent, respectively.

Benefits were similar between the overall population, in the patients who had received bevacizumab and those who did not, and in those with PD-L1 CPS of ≥1 as well as those with CPS ≥10. At longer follow up of a median of 39 months, addition of pembrolizumab to chemotherapy improved median OS in the overall group (26 versus 17 months), among those with CPS ≥1 (29 versus 17 months), and among those with CPS ≥10 (30 versus 17 months) [32]. However, only a small number of patients had CPS <1 (69 patients), and it remains unclear if this group benefits from the addition of pembrolizumab (HR for OS 0.87, 95% CI 0.50-1.52).

The most common grade 3 to 5 adverse events were anemia (30 percent in the pembrolizumab group and 27 percent in the placebo group) and neutropenia (12 and 10 percent, respectively). Toxicities of immune checkpoint inhibitors are discussed in more detail elsewhere. (See "Toxicities associated with immune checkpoint inhibitors".)

Based on these data, pembrolizumab is approved by the US Food and Drug Administration (FDA) in combination with chemotherapy, with or without bevacizumab, for the treatment of patients with persistent, recurrent, or metastatic cervical cancer whose tumors express PD-L1 (CPS ≥1) [33].

Second-line therapy — For women who have progressed after first-line treatment and those patients who are not candidates for combination chemotherapy, we suggest single-agent therapy. However, there is no evidence that treatment in the second- or later-line setting improves OS compared with best supportive care in this population.

A choice among active agents must be tailored to the individual patient, with consideration to prior therapies received, residual toxicity, and performance status. For those who have progressed on prior chemotherapy for metastatic disease and desire further treatment, we assess tumor PD-L1 expression on existing tissue for consideration of pembrolizumab (or consider repeat biopsy, if tissue is not available). For women whose tumors display PD-L1 expression, pembrolizumab may reasonably be offered before additional cytotoxic therapy, provided it was not used in the front-line setting. Given the limited activity of currently available agents, we encourage participation in well-designed clinical trials exploring alternative approaches to metastatic cervical cancer.

The most active single agents are:

●Carboplatin – ORR 15 percent [34].

●Paclitaxel (175 mg/m² intravenous [IV] every three weeks with dose reduction to 135 mg/m² if patients received prior RT) – ORR 20 to 25 percent [35-37].

●Topotecan (1.5 mg/m² IV daily for five days every 21 days) – ORR 19 percent [38-41].

●Tisotumab (2 mg/kg IV every three weeks) – Tisotumab vedotin-tftv is a tissue factor-directed antibody and microtubule inhibitor drug conjugate approved by the US Food and Drug Administration (FDA) for recurrent or metastatic cervical cancer that has progressed on chemotherapy [42]. In a single-arm trial that treated 101 patients with recurrent or metastatic cervical cancer who had received no more than two prior systemic regimens in the recurrent or metastatic setting, including at least one prior platinum-based chemotherapy regimen, the ORR to tisotumab was 24 percent (17 percent partial response and 7 percent complete response) [43]. The most common treatment-related adverse events included alopecia (38 percent), epistaxis (30 percent), nausea (27 percent), conjunctivitis (26 percent), fatigue (26 percent), and dry eye (23 percent). Grade ≥3 treatment-related adverse events were reported in 28 percent of patients and included neutropenia (3 percent), fatigue (2 percent), ulcerative keratitis (2 percent), and peripheral neuropathies (2 percent each with sensory, motor, sensorimotor, and neuropathy peripheral). (See "Ocular side effects of systemically administered chemotherapy", section on 'Antibody-drug conjugates'.)

●Checkpoint inhibitors – Immune checkpoint inhibitors are an option for those who have not received them in a prior line of treatment. Pembrolizumab (200 mg IV every three weeks) has a role in the first-line setting for metastatic cervical cancers, but it may be used in the second-line setting for those with PD-L1 CPS ≥1 percent, per the FDA [44]. Cemiplimab (350 mg IV every three weeks) has also demonstrated efficacy in platinum-resistant recurrent cervical cancer that was not previously treated with immunotherapy (irrespective of PD-L1 expression). It does not have regulatory approval for cervical cancer, however. (See 'Addition of a checkpoint inhibitor' above.)

•For cervical cancers that were not treated initially with pembrolizumab, it has shown promise as second-line therapy in PD-L1-positive cancers [44]. For example, in a phase II basket study of pembrolizumab, among 82 patients with advanced, pretreated cervical cancer and PD-L1 expression of 1 percent or more, the ORR was 15 percent [45]. Similar findings were observed in the KEYNOTE-028 trial, in which the ORR was 17 percent and duration of response was 5.4 months [46]. Responses in tumors with high mutational burden are discussed in detail elsewhere. (See "Tissue-agnostic cancer therapy: DNA mismatch repair deficiency, tumor mutational burden, and response to immune checkpoint blockade in solid tumors", section on 'Tumors with high mutational burden'.)

•In a randomized trial in patients with disease progression after platinum-based chemotherapy, cemiplimab improved OS relative to single-agent chemotherapy (12.0 versus 8.5 months; HR 0.69, 95% CI 0.56-0.84), as well as objective response rates (16 versus 6 percent) [47]. Cemiplimab versus chemotherapy resulted in a median PFS of 2.8 and 2.9 months, respectively, but cemiplimab increased duration of response (16.4 versus 6.9 months).

PD-L1 expression was assessed using the SP263 monoclonal antibody (Ventana). Patients with cancers that were PD-L1 positive (expression on ≥1 percent of tumor cells) experienced similar clinical outcomes as the overall population. Grade ≥3 events occurred in 45 percent in the cemiplimab group and 53 percent in the chemotherapy group.

Toxicities of checkpoint inhibitors include rash, colitis, and pneumonitis, among others, and are discussed in detail elsewhere. (See "Toxicities associated with immune checkpoint inhibitors".)

Other less frequently used options include:

●Nanoparticle, albumin-bound paclitaxel (125 mg/m² on days 1, 8, and 15 every 28 days) – ORR 29 percent [48].

●Vinorelbine (30 mg/m² IV push weekly for two weeks every 21 days) – ORR 15 percent [49-51].

●Pemetrexed (900 mg/m² IV every three weeks) – ORR 15 percent [52,53].

●Ifosfamide (1.2 g/m² IV daily for five days every 28 days) – ORR 22 percent [38,54].

●Irinotecan (350 mg/m² IV every three weeks, or 125 mg/m² weekly for four weeks followed by a two-week washout period) – ORR 15 percent [55,56].

MANAGEMENT OF ACUTELY SYMPTOMATIC PATIENTS — For women with pelvic pain or bleeding from locally advanced disease, and for patients with symptomatic metastatic disease, single-dose or short-course external beam radiation therapy (RT) may be useful. However, treatment planning is critical for women who have previously been treated with RT due to the risks of excessive toxicity to normal tissue [57-60]. For women with symptoms from a vaginal or pelvic recurrence, the optimal regimen has not been established. A typical regimen consists of twice daily fractionations of external beam RT (3.7 Gy per fraction) for two consecutive days repeated at four-week intervals for three courses [57,60].

For the treatment of visceral or distant metastases, large fractions of RT administered in a short treatment course may also be used for local treatment of symptoms related to metastatic disease, such as alleviation of pain from skeletal metastases, or symptoms associated with brain lesions. (See "Radiation therapy for the management of painful bone metastases" and "Epidemiology, clinical manifestations, and diagnosis of brain metastases".)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Treatment of cervical cancer".)

SUMMARY AND RECOMMENDATIONS

●Introduction – Following treatment of early-stage cervical cancer, distant metastases or multiple recurrence sites develop in 15 to 61 percent of cases, usually within the first two years of completing treatment. (See 'Introduction' above.)

●Clinical presentation – Recurrent cervical cancer presents as disease isolated to the pelvis (locoregional recurrence) or with disease involving other organs or outside the pelvis. If a vaginal recurrence is suspected, the area of concern should be biopsied to prove recurrent disease. (See 'Clinical presentation' above.)

All patients suspected of recurrent disease should undergo positron emission tomography (PET)/computed tomography (CT) for evaluation of local and distant disease. (See 'Clinical presentation' above.)

●Management of local recurrence – For patients with a local recurrence, we suggest surgical resection, rather than nonsurgical approaches, if they are appropriate surgical candidates based on tumor recurrence, age, and comorbidities (Grade 2C). In select patients, surgical resection may present a curative option. However, for patients with a local recurrence who have not received radiation therapy (RT), RT in combination with chemotherapy is an acceptable alternative, and is preferred in those who are not surgical candidates, provided they have not previously undergone pelvic or intravaginal RT. (See 'Management of local recurrence' above.)

●Disease limited to the nodes – For women presenting with disease limited to the nodes (ie, para-aortic and/or supraclavicular nodes), some experts prefer to treat with systemic chemotherapy. Others prefer to administer RT (with or without chemotherapy). A choice between them depends on institutional practice and patient preferences.

●Isolated metastatic disease – For women presenting with isolated metastatic disease (ie, to the lung or liver), we suggest surgical resection if they are appropriate candidates (Grade 2C). These patients may achieve a sustained clinical remission following resection of disease.

●Other metastatic disease – For women with recurrent cervical cancer, those who are not surgical candidates, and those who present with metastatic disease, our approach is as follows:

•We suggest first-line treatment with chemotherapy plus bevacizumab rather than chemotherapy alone (Grade 2B). However, the costs of therapy may require scrutiny in comparison to the benefits and risks of incorporating bevacizumab in this setting, especially in low-resource settings. (See 'Initial treatment' above.)

•We also recommend the addition of either atezolizumab (irrespective of programmed cell death ligand 1 [PD-L1] status) or pembrolizumab (for PD-L1 positive disease only) (Grade 1B). The choice between atezolizumab and pembrolizumab in PD-L1 positive disease should be made based on cost and availability, given a lack of direct comparisons of these two agents in this setting. (See 'Addition of a checkpoint inhibitor' above.)

•We suggest a platinum-based combination rather than a nonplatinum-based regimen or single-agent cisplatin (Grade 2B). We suggest cisplatin plus paclitaxel rather than other platinum-based regimens (Grade 2C). Carboplatin plus paclitaxel is a reasonable alternative, particularly for patients who are poor candidates for cisplatin (eg, due to pre-existing renal failure) and patients previously treated with cisplatin-based chemoradiation. (See 'Initial treatment' above.)

●Second-line therapy – For patients who progress following first-line chemotherapy and those who are not candidates for combination therapy, we suggest single-agent therapy. A choice among active agents must be tailored to the individual patient, with consideration to prior therapies received, residual toxicity, and performance status. (See 'Second-line therapy' above.)

●Special consideration for vaginal metastases – For patients who are symptomatic due to vaginal metastases or other lesions, we suggest RT for palliation (Grade 2C). (See 'Management of acutely symptomatic patients' above.)