Efficacy and Safety of Immunotherapy for Cervical Cancer—A Systematic Review of Clinical Trials

Simple Summary Cervical cancer is the 4th leading cause of cancer deaths in women worldwide. Surgery, chemotherapy, radiotherapy and chemoradiation therapy are routinely used in the treatment of cervical cancer, while immunotherapy remains a novelty. The aim of our systematic review was to provide an extensive overview of the efficacy and safety of immunotherapy in cervical cancer patients. A total of 50 clinical trials assessed immune checkpoint inhibitors, therapeutic vaccines and adaptive cell transfer therapy. Overall, immunotherapy showed an acceptable safety profile. While the level of evidence on efficacy is still low, promising results, including few complete remissions in heavily pretreated women with metastatic disease, have been observed. Furthermore, a recent phase III trial assessing pembrolizumab in combination with chemotherapy (±bevacizumab) demonstrated a prolonged overall survival and has now led to a new standard of care for first-line systemic treatment in persistent, metastatic or recurrent cervical cancer patients. Abstract Purpose: To systematically review the current body of evidence on the efficacy and safety of immunotherapy for cervical cancer (CC). Material and Methods: Medline, the Cochrane Central Register of Controlled Trials and Web of Science were searched for prospective trials assessing immunotherapy in CC patients in compliance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Full-text articles in English and German reporting outcomes of survival, response rates or safety were eligible. Results: Of 4655 screened studies, 51 were included (immune checkpoint inhibitors (ICI) n=20; therapeutic vaccines n = 25; adoptive cell transfer therapy n=9). Of these, one qualified as a phase III randomized controlled trial and demonstrated increased overall survival following treatment with pembrolizumab, chemotherapy and bevacizumab. A minority of studies included a control group (n = 7) or more than 50 patients (n = 15). Overall, response rates were low to moderate. No response to ICIs was seen in PD-L1 negative patients. However, few remarkable results were achieved in heavily pretreated patients. There were no safety concerns in any of the included studies. Conclusion: Strong evidence on the efficacy of strategies to treat recurrent or metastatic cervical cancer is currently limited to pembrolizumab in combination with chemotherapy and bevacizumab, which substantiates an urgent need for large confirmatory trials on alternative immunotherapies. Overall, there is sound evidence on the safety of immunotherapy in CC.


Introduction
Cervical cancer is the 4th most common cancer type in women and the most common gynecological tumor, accounting for around 342,000 deaths in 2020 [1]. Primary treatment options include surgery, (chemo)radiation therapy ((C)RT) or systemic chemotherapy (CHT) [2]. While great advances in preventing cervical cancer by prophylactic vaccinations have been achieved, systemic treatment options, especially for advanced, metastatic or recurrent cervical cancer, are still limited [3].
In 2017, the anti-angiogenesis drug Bevacizumab led to prolonged survival rates of around 3.5 months when combined with CHT [4]. It has since become the treatment of choice for primary therapy of persistent, metastatic or recurrent disease. However, 5-year recurrence rates remain high .6% for stage IIB-IVB), and patients being treated with 2nd line systemic therapy are faced with a mean overall survival time of around 7-9 months [5][6][7]. To date, no clear superiority of any ≥2nd line CHT or targeted therapy has been demonstrated upon recurrence [2]. Furthermore, substantial side effects of CHT need to be considered, especially in elderly multimorbid patients, and palliative care is a viable option that has to be discussed given the lack of effective systemic treatments in these patients.
With its success in the treatment of lung cancer, melanoma or renal cell carcinoma, immunotherapy has gained increasing popularity in recent years. The National Cancer Institute defines immunotherapy as a "type of therapy that uses substances to stimulate or suppress the immune system to help the body fight cancer, infection or disease" [8] and includes different approaches such as immune checkpoint inhibitors (ICI), adaptive cell transfer therapy (ACTT), therapeutic vaccines and immune system modulators. Overall, ICIs are currently the most prominent representatives of immunotherapy and are being investigated in numerous cancer types, including gynecological cancers [9]. However, as most cervical cancers are associated with the human papillomavirus (HPV), targeted immunotherapies such as therapeutic vaccines or ACTT are also emerging. While immunotherapy is already an integral part of therapy in some cancer types, tumor responses to immunotherapy can vary drastically between cancer types. Despite various promising approaches, immunotherapy is still at the beginning of being clinically explored for cervical cancer.
Thus, this systematic review aims to present the current clinical evidence on the efficacy and safety of immunotherapy in cervical cancer patients, mainly addressing the following questions: (1) Which immunotherapies have been clinically assessed in cervical cancer patients?; (2) Does cervical cancer respond to immunotherapy treatment?; (3) Does immunotherapy prolong survival in cervical cancer patients? As a secondary aim, the safety of immunotherapy in cervical cancer patients is evaluated.

Materials and Methods
This systematic review is reported in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [10]. The systematic review was prospectively submitted to PROSPERO. Due to the automatic check currently performed by PROSPERO the submission was rejected just after finishing the manuscript because of an incorrectly filled question. Thus, a correction was not possible anymore and no registration number is available.

Literature Search
Three electronic bibliographical databases, MEDLINE (via Ovid), the Cochrane Central Register of Controlled Trials (CENTRAL) and Web of Science, were searched systematically without any restrictions towards language or publication date [11]. The creation and optimization of the search strategies for each database were aided by a librarian from Johannes-Gutenberg University Mainz. The search strategies were developed according to the following PICOS criteria [12]: • I (intervention)-Any form of immunotherapy; • C (comparison)-Any (including chemotherapy, targeted therapy, surgery or placebo) or no comparison; • O (outcome)-At least one measure of survival outcomes, response rates or adverse events; • S (study design)-All types of prospective study designs.
All search strategies included index terms as well as free text related to cervical cancer and immunotherapy, including therapeutic vaccines, checkpoint inhibitors and CAR T cells. The search strategies are provided in the Supplementary Material (Appendix A). The search was performed on the 29th of September 2021. A cross-reference check was performed on all included studies by screening their reference lists and by using Google Scholar to identify articles that cite the included studies. Furthermore, studies included in related systematic reviews and meta-analyses were screened for eligibility. This was performed between the 23rd-30th of November 2021. Grey literature, including conference abstracts or commentaries, were not considered for the systematic review; however, highly relevant abstracts were included in the discussion if sufficient data was provided.

Eligibility Criteria
Original, prospective clinical trials (phase I-IV) published as complete journal articles were included. Eligibility criteria were based on the PICOS criteria reported above. Furthermore, only full-text articles written in English or German were included. Studies were excluded if (1) only a single patient with cervical cancer was reported (e.g., case reports, studies recruiting various cancer types), (2) a more recent publication of the same study was available (except when reporting different outcomes) or (3) due to a retrospective study design. As single-arm cohort trials are commonly used in phase I and II clinical trials in oncology and only a few comparative trials were expected in this emerging field, non-comparative trials were also included in this systematic review, despite the known increased risk of bias in these trials. Results were and should be interpreted accordingly.

Study Selection and Data Extraction
Title and abstract screening, as well as full-text screening, were conducted by two review authors (M.W.S. and K.A.) independently. Any disagreements over a particular study were resolved through discussion with a third reviewer (M.J.B.). Data extraction was performed by M.W.S. and re-checked independently by K.A. using predefined word spreadsheets, which were tested and adapted based on a few sample studies. Data extraction included an individual study identifier (author, title, and year of publication), fundamental study details including study population, study phase, programmed death ligand 1 (PD-L1) status for ICI trials and HPV status for therapeutic vaccines and ACTT, interventions and results. Efficacy data were extracted as survival and response data. Survival data included: (1) months and confidence interval (CI) for progression-free survival (PFS) and overall survival (OS), (2) (estimated) PFS or OS rates at 1 year and/or longest follow-up in percent, and (3) recurrence rates (%) for patients treated with curative intent. Response parameters included: (1) objective response rates (ORR, %), (2) disease control rates (OS, %) and (3) duration of response (DOR, months). Time to response was not extracted as initially planned, as it was reported by too few trials. When not provided, ORR (complete response + partial response) and DCR (complete response + partial response + stable disease) were calculated with available data. When possible, data for the subgroups based on PD-L1 status were reported separately.
Safety details of treatment-related adverse events (TRAE) were extracted as percentages based on the following: (1) treatment-related deaths (TRD), (2) TRAE of any grade (%), (3) TRAE grade 3 or higher (%), (4) list of TRAEs occurring in more than 5% (TRAEs≥ grade 3) or 10% (TRAEs of any grade), and (5) if available, the overall percentage of potentially immune-mediated AEs or the list of potentially immune-mediated AEs above 10% if the overall percentage was not given. If TRAEs were not reported, adverse events (AE) were described following the same rules.

Risk of Bias Assessment
Two review authors (M.W.S. and K.A.) independently assessed the risk of bias and study quality, and disputes were settled in discussion with a third reviewer (M.J.B). The revised Cochrane Risk of Bias tool for randomized trials (RoB-2) [13] was used for randomized control trials (RCTs). Unfortunately, no adequate risk of bias assessment tool with high validity evidence exists for single-arm cohort trials, which includes the majority of phase I and II trials in oncology, and available tools for comparative trials are not applicable to non-comparative trials. Thus, the risk of bias in single-arm cohort trials was assessed to judge the reliability of the outcome of each study based on domains assessed by the Risk of Bias in Non-Randomised Studies of Interventions tool (ROBINS-I tool) [14]. The risk of bias was judged as high, low or unclear. As no direct comparison between groups was made, the item "bias in selection of participants" was not evaluated. For bias due to confounding, factors were judged that may have an effect on the main outcomes, such as the use of immune-suppressants or other potential treatments upon the termination of trial treatments. Bias due to deviations from intended interventions was considered high for any other analysis than intention-to-treat, due to the possibility of overestimating the true efficacy benefit. To assess bias due to missing data, an availability of 90% (due to small sample sizes, one missing person can already be below the often used 95% threshold) of outcome data was regarded as sufficient. The risk of bias in measurement outcomes was judged based on the objectivity of efficacy or safety data, meaning the likelihood of misclassifying an outcome (e.g., stable vs. progressive disease in clinical examinations). Bias in the classification of interventions and bias in the selection of reported results was judged as written in the ROBINS-I tool. A similar approach of adapting the ROBINS-I tool was recently used in a Cochrane systematic review by Jullien et al. [15]. The risk of bias was assessed on the study level based on efficacy outcomes. For studies in which no adequate efficacy measurements were assessed, risk of bias judgment was done for clinical safety evaluations.

Study Selection
Through the search, a total of 4655 studies were identified, of which 51 were included in the systematic review. A detailed flowchart of the study selection process is depicted in Figure 1.

Efficacy of Immune Checkpoint Inhibitors Based on PD-L1 Status in Cervical Cancer
Four studies with monotherapy of ICIs or placebo control reported subgroup analyses with regards to PD-L1 status in cervical cancer patients. Colombo et al. [17] reported a better HR with regards to PFS in PD-L1 positive patients treated with pembrolizumab plus chemotherapy ± bevacizumab compared to PD-L1 negative patients (CPS 1-<10% 0. . On the contrary, a response rate of 7.9% in PD-L1 negative patients (n = 38%) was seen under balstilimab monotherapy in patients with metastatic, persistent or recurrent cervical cancer. However, a larger ORR of 20% was seen in PD-L1 positive patients (n = 58) [34].

Efficacy and Safety of Checkpoint Inhibitors with or after (Chemo)Radiotherapy
When considering the use of ICIs in first-line therapy, their safety and efficacy combined with CRT are important to evaluate. Duska et al. conducted a phase II RCT to assess the safety of pembrolizumab during vs. after CRT in 52 locally advanced cervical cancer patients [20]. In their preliminary report, an overall similar and acceptable safety profile was demonstrated. The study is currently ongoing, and a follow-up including long-term safety data is expected. Mayadev et al. evaluated the use of 4 cycles of ipilimumab after CRT with an acceptable safety profile. Median PFS and OS were not reached within the 14.8-month follow-up period [30]. The combination of ICI and RT in patients resistant or intolerant to platinum and taxane doublet chemotherapy was assessed by Rischin et al. by administering cemiplimab with or without concurrent hypofractionated RT in a nonrandomized controlled trial. Most included patients (90%) had received prior cancer-related RT as well as bevacizumab (70%). One patient died due to severe pneumonitis in the concurrent hypofractionated RT. In general, a similar safety profile was observed [33]. The overall response rates were comparable between both groups, with an increased DCR in the hypofractionated RT group (60% [95% CI 26.2-87.8] vs. 40% [95% CI [12.2-73.8]), but an increased median OS in the cemiblimab monotherapy group (10.3 months vs. 8.0 months) as well as an increased DOR (11.2 months vs. 6.4 months). However, the upper limit of the 95% CI was not reached in both groups.

Therapeutic Vaccines
Human papilloma virus (HPV) infections are known to play an important role in the etiology of the majority of cervical cancer cases. While prophylactic vaccinations against HPV are widely available and have been proven to be effective in preventing cervical cancer [38], they are unable to eliminate existing tumor cells and precursor lesions. Thus, there is still a major need for therapeutic vaccines which aim to evoke a durable and strong immune response eliminating cancer cells carrying special antigens. Tumor antigens can roughly be divided into tumor-associated antigens, which can also be found on healthy cells but are generally overexpressed in tumor cells, and tumor-specific antigens, such as oncogenic viral antigens, which are foreign to the healthy cell [39]. E6 and E7 antigens are known to cause HPV-associated neoplastic changes by allowing the uncontrolled progression of cell cycles into the S phase [40]. Because many cervical cancers are associated with HPV, these tumor-specific antigens present a promising target to provoke specific immune responses without increasing autoimmunity.

Clinical Vaccine Trials in Cervical Cancer
A total of 25 trials were identified assessing therapeutic vaccines in cervical cancer patients [21,, of which 4 were not specific to cervical cancer but included a larger population with various cancers. The trials were published between 1989 and 2020. An overview can be found in Table 3. To date, no phase III trial has been published. Three trials assessed the combination of therapeutic vaccines with concomitant chemotherapy (carboplatin/paclitaxel q3w). Two of them administered the vaccinations two weeks after chemotherapy starting with the second cycle [43,51], while Basu et al. interrupted the vaccinations for five weekly cycles of cisplatin [62]. Only Youn et al. evaluated the combined use of ICIs (pembrolizumab) with a therapeutic vaccine [21]. Vaccines were mostly injected subcutaneously; however some were given intravenously, intradermally or intramuscularly. They ranged from single-shot doses, a predefined number of doses to unlimited doses, or repeated injections until disease progression. No deaths were reported in immediate relation to the vaccine therapies. Common treatment-related adverse events included injection site reactions such as swelling, redness, itching, pain as well as systemic reactions with fever or flu-like symptoms. Overall only a few grade III or IV toxicities occurred.

Adoptive T Cell Transfer Therapy
Another interesting approach is adoptive cell transfer therapy, such as the use of cytokine-induced killer cells (CIK) or T cell transfer. The two main approaches of adoptive T cell therapy include the use of tumor-infiltrating lymphocytes (TILs) and engineered T cell receptor (TCR)/chimeric antigen receptor (CAR) T cell therapy, for which tumor-specific autologous or allogenic T cells are grown ex vivo and reinfused for treatment [65]. TILs do not have to be modified as they are gained from tumor biopsies of patients and are thus expected to recognize tumor-associated antigens. They are grown in the laboratory with the help of interleukins before being reinfused. Patients usually have to undergo lymphodepletion by CHT or RT. CAR T cell therapy, on the other hand, is an example of genetically modified T cells to express CARs. They do not rely on major histocompatibility complexes (MHC) to present tumor antigens that are often downregulated in tumor cells [66] but can directly recognize surface antigens of tumor cells. Another alternative to CAR T cell therapy but with a similar approach is modifying the physiological T cell receptor (TCR) complex to recognize specific tumor antigens.

Clinical Trials in Cervical Cancer Patients
A total of 9 clinical trials evaluated adoptive cell transfer therapies using CIKs (n = 3, [23,67,68]), TILs (n = 3, [27,69,70]) or engineered TCR (n = 3, [71][72][73]) in a total of 209 cervical cancer patients between the years 2015-2021. An overview of all trials can be found in Table 4. Up to now, CAR T cell therapy has not been clinically assessed for cervical cancer. Three trials reported treatment with TCR in a broader cancer population, including cervical cancer [71][72][73], and one trial evaluated combined immunotherapy with TILs and anti-PD1 (nivolumab) [27]. In all trials, except for the combination with nivolumab, chemotherapy was given prior to or during adoptive cell therapy. Five studies assessing TCR-engineered T cells or TILs used cyclophosphamid and fludarabine, a non-myoablative chemotherapy [69][70][71][72][73]. The other two larger trials were designed as RCTs assessing CHT vs. CHT plus dentric cell-cytokine induced killer (DC-CIK) cell infusions [66] and CHT plus RT vs. CHT plus RT and CIK infusions [68]. While Chen et al. demonstrated significantly prolonged survival rates through the addition of DC-CIK ACTT, Li et al. did not find the same difference with CIK ACTT combined with CHT and RT. However, the study population differed between both trials.
Overall, adoptive cell transfer therapy was proven to be safe in cervical cancer patients, with most observed adverse events being due to the associated chemotherapies and highdose interleukin treatments. No autoimmune reactions were observed.

Nonspecific Immune System Modulators/Immunomodulating Agents
Various nonspecific immune system modulators, including interferon α, interleukin 12, extract derived from agaricus blazei murill kyowa, mycobacterium tuberculosis (Z-100) or streptococcal preparations (OK-432) as well as sizofiran and thymopentin or cornyebacterium parvum have been studied for cervical cancer . All agents affect the immune system in a general way at different levels. However, probably due to the rise of targeted immunotherapies, only a few large trials assessing nonspecific immune system modulators have been conducted after the year 2000. Unfortunately, the term immunotherapy (including variations of it) was not as commonly used to describe a treatment approach back then, and our search strategies were found to be inappropriate to adequately report the current evidence of nonspecific immune system modulators in this systematic review. Thus, the respective trials were excluded. Nevertheless, a sample of 25 trials can be found in Table S1.

Risk of Bias Assessment
Only one trial was judged to be at low risk of bias ( Figures S1-S4). In RCTs, the risk of bias was mostly high due to a high number of patients lost to follow up or dropping out as well as the non-blinding of patients and outcome assessors. All non-comparative trials were judged as at least unclear risk of bias, as confounding factors were not reported or regarded in these trials. Furthermore, missing outcome data, as well as inappropriateness of recurrence or response detection in trials conducted around the year 2000 (only clinical examinations or x-rays), were the most common reasons for a high-risk judgment in noncomparative trials. One study reported two different groups with varying therapies, but no comparison was made between these groups [33]. The study was thus assessed as a non-comparative trial.

Discussion
Despite constant advances in cancer therapy, the prognosis of locally advanced, recurrent or metastasized cervical cancer patients remains unsatisfactory. With a better understanding of tumor immunology, including tumor mechanisms of resistance and avoidance to the host's immune response, immunotherapy has become one of the most promising approaches in cancer treatment. The idea of increasing and targeting the body's already occurring natural fight against aberrant cells seems simple, but finding the most effective approach is difficult.
The highest level of evidence for immunotherapy is currently available for ICIs (Tables 1 and 2). Based on the positive results of phase I/II trials on pembrolizumab for cervical cancer, the FDA approved pembrolizumab as a monotherapy in patients with recurrent or metastatic PD-L1 positive disease in 2018. Due to the recently published phase III trials by Colombo et al. with a prolonged OS of around 8 months, pembrolizumab has now been approved in patients in combination with CHT and bevacizumab [17]. While this is currently the only phase III trial for ICIs in cervical cancer, many are ongoing, and results are eagerly awaited. Promising interim results of the phase III EMPOWER trial were presented at the European Society for Medical Oncology (ESMO) congress in 2021. Recurrent or metastatic cervical cancer patients who had progressed after platinum-based CHT were treated with either cemiplimab or the investigators' choice of CHT. An interim analysis of 608 patients clearly favored cemiplimab treatment with regards to OS (12.0 vs. 8.5 months, p < 0.001), PFS and ORR [98]. In summary, the current evidence on ICI for cervical cancer is encouraging; however, ORR for ICI monotherapy in patients progressing after platinum-based chemotherapy is still low. In particular, the subgroup of PD-L1 negative patients does not seem to benefit from ICI in cervical cancer. Considering the ORR of around or less than 25% for ICI monotherapy in cervical cancer, several studies aim to identify prognostic factors to anticipate a favorable reaction to ICIs in cancer patients [99]. Alternative approaches try to increase response rates and to avoid acquired immune resistance by combining ICIs with other systemic therapies. For example, by combining a therapeutic vaccine with pembrolizumab (anti-PD-1), a remarkable ORR of 42% and a DCR of 58% was reached in advanced or recurrent HPV-positive cervical cancer patients [21]. Even higher response rates of 55% and a DCR of 82% were reported from the CLAP trial, treating patients with camrelizumab (anti-PD1) and apatinib (a tyrosine kinase inhibitor), despite more than 55% of the patients having had 2 prior lines of systemic chemotherapy [32]. Interim results of the CheckMate 358 study demonstrated the efficacy of 2 combinations of nivolumab (anti-PD1) and ipilimumab (anti-CTLA4) in patients with recurrent and metastatic cervical cancer. Results were presented at the 44th European Society for Medical Oncology (ESMO) congress in 2019 [100]. However, between-study comparisons should be interpreted with caution, and direct comparisons of treatment regimens are needed to prove the superiority of either combination. Another interesting approach to improve the efficacy of ICI is to increase the amount of PD-L1 on the cell surface of tumors. Recently, the extracellular plasminogen activator inhibitor type I (PAI-1) was found to be responsible for internalizing PD-L1, and targeting PAI-1 with a pharmacological inhibitor (tiplaxtinin) has led to increased PD-L1 expression on tumor cells in vivo and in vitro. Thus, a combination of ICI with tiplaxtinin seems promising and has shown first synergistic effects in a murine model of melanoma [101]. Overall, ICI treatment in cervical cancer proved to be safe with expected adverse events observed in other cancer types [102]. The occurrence of adverse events slightly increased with the combination of ICI and CHT [17]. Nevertheless, no increase in severe toxicities was observed in combinations with CHT or RT. Various trials assessing ICI are currently ongoing, including combinations with RT or CRT (nivolumab (NCT03298893/NCT03527264), atezolizumab (NCT03612791/NCT03612791), dostarlimab (NCT03833479), durvalumab (NCT03830866), pembrolizumab (NCT04221945/NCT02635360), as well as evaluating ICI for neoadjuvant CHT (pembrolizumab (NCT04238988)).
Despite major advances in the use of prophylactic HPV vaccines, therapeutic vaccines for HPV+ or HPV-cervical cancers are still at the beginning of their development and use in humans. This is clearly demonstrated by the limited amount of phase II trials, as described in this systematic review. Nevertheless, the concept of therapeutic vaccinations to fight cancer is of great interest and has shown promising results in many pre-clinical trials [103]. Due to the large proportion of HPV infections in cervical cancers, the HPV oncogenes E6 and E7 are the targets of the majority of the tested therapeutic vaccines for cervical cancer. Overall, therapeutic vaccinations have proven to be safe in numerous phase I and II clinical trials (Table 3). No major allergic reactions occurred in the here-reported trials. Even in combination with chemotherapy [43,51] or with PD-1 checkpoint inhibitors [21], therapeutic vaccinations did not lead to a notable increase in serious adverse events. Interestingly, the immunological T cell response was stronger when vaccinations were given during chemotherapy rather than post-chemotherapy. T cell reactivity around two weeks after chemotherapy was found to be increased after the second cycle of chemotherapy and subsequent ones, possibly due to the normalization of abnormally high tumor-promoting myeloid cell populations, which are initially higher in the presence of a large tumor burden [51]. Similar immunological changes in response to chemotherapy have been demonstrated in ovarian cancer patients [104,105]. Due to the lack of RCTs, no concluding statement on the efficacy of the currently tested therapeutic vaccines can be given. However, the majority of therapeutic vaccinations were able to demonstrate an immunological response in cervical cancer patients, which may prolong OS. Melief et al. reported high ORR and DCR of 43% and 86%, respectively, and found significantly improved OS in a group of 77 cervical cancer patients with strong (higher than the median) immunological vaccine responses compared to those with low (lower than the median) immunological vaccine responses (16.8 months vs. 11.2 months; p = 0.012) when treated with an HPV E6 and E7 peptide-based vaccine ± pegylated INFα in addition to chemotherapy [43]. A meta-analysis found the OS in similar patient populations (advanced, metastatic or recurrent cervical cancer) treated with chemotherapy alone to be around 10-12.8 months [106]. Promising interim results of a currently ongoing phase II trial assessing a triple combination of an HPV 16 E6/7 based vaccine, a tumor-targeting IL12 immunocytocine and bintrafusp alfa (a PD-L1 and TGF-ß inhibitor) in patients with advanced, previously treated cervical, anal, head and neck, vulvar and vaginal cancer (n = 25) were recently presented at the ASCO Annual Meeting 2021. The triple therapy led to an ORR of 55.6% with an ongoing response of 80% after 8 months of follow-up. All checkpoint-inhibitor-naïve patients (n = 6) are still alive [107]. To date, there is no data on the safety and efficacy of combining therapeutic vaccines with CRT. However, an ongoing trial IMMNUOCERV (NCT04580771) is currently assessing a liposomal HPV-16 E6/E7 multipeptide vaccine (PDS0101) combined with CRT in advanced cervical cancer patients. Overall, therapeutic vaccines are a growing and promising research trend in cancer therapy, with multiple ongoing clinical trials, especially for HPV-associated cancers, including head and neck cancers, cervical, vulva, vaginal or anal cancers [108].
ACTT can be an appealing alternative strategy to target tumor-specific antigens and has shown remarkable results with complete responses in some patients with breast cancer [109] or metastatic melanoma [110]. Presently, few clinical trials have been conducted in cervical cancer patients, although with promising results. A total of 3 phase I and II trials on ACTT demonstrated complete remissions in altogether 5 pretreated, metastatic cervical cancer patients ongoing at 15-67 months. However, ORR was still only around 28-33% in these relatively small study populations (a total of 30 patients in these 3 studies) [69,70,72]. A phase I trial with antigen receptor-engineered T cell therapy against HPV E7 showed anti-tumor efficacy even in cervical and other cancer patients pretreated with PD-1 based immunotherapy. The authors explained this through the contrasting mechanism of actions. While TCR-engineered T cells directly target the tumor, PD-L1 checkpoint inhibition acts by disinhibiting the physiological anti-tumor response [73]. Based on this rationale, one could expect a positive effect by combining adoptive cell therapy with anti PD1 immunotherapy. As shown by Yin et al., combining TILs with nivolumab led to a response rate of 25% even in PD-L1 negative patients [27]. However, whether this is due to the combination or the TILs alone cannot be determined. Whether or not the concurrent single cycle of lymphocyte-depleting CHT adds to the antitumor effects of the treatment is unclear. Although cyclophosphamide has demonstrated antitumoral effects in several malignancies, it is not used in cervical cancer treatment. However, its analog ifosfamide has shown low response rates of 15.7% in platinum-naïve and 11% in platinum-treated cervical cancer patients with short durations of response ranging from 1.8-3.1 months [111,112]. Other than using HPV-targeted ACCT, the use of CIK has been explored for cervical cancer. Chen et al. found a significantly prolonged OS (3 year OS rates: 56.4% vs. 80%) as well as decreased recurrence rates (3-year recurrence rates: 46.2% vs. 22.5%) in cervical cancer patients stage IIB-IV post-surgery when treated with DC-CIK in addition to CHT with cisplatin in an RCT. The tolerability of the treatment regimen was not reported [67]. On the other hand, Li et al. found no significant difference with regard to OS in a mostly pretreated advanced cervical cancer cohort (Stage IIA-IV) when treated with CIK in addition to RT and CHT in their RCT, despite a significantly increased short-term ORR after 1 month (88.6% with CIK vs. 68.9% control, p < 0.05). However, the randomization process was not blinded, leading to an overall high risk of bias in this trial [68]. One explanation for the differing results could be the lack of co-culturing of DC with the CIK by Li et al., as it was shown that the cytotoxic abilities of CIK can be enhanced by co-culturing and resulting stimulation by DC [113]. However, the effect of both CIK and DC-CIK was demonstrated, for example, in lung cancer [114], gastric cancer [115] and colon cancer [116]. Thus, further clinical trials are warranted in cervical cancer based on these promising results. Currently, ongoing trials include a multicenter phase II trial assessing the efficacy and safety of TIL ± pembrolizumab (NCT03108495), a phase II trial of CIK in addition to radiofrequency (NCT02490748), a phase II trial for T cell receptor gene therapy targeting HPV 16 E7 in HPV-associated cancers (NCT02858310) and a phase I trial on HPV-E6 specific TCR-T cells ± anti PD1 therapy (NCT03578406).
The major limitations of this systematic review include the small number of highquality trials, especially the lack of RCTs, as well as the heterogeneity in study populations making direct comparisons of trial results unreliable, which is why no data synthesis has been performed. Furthermore, in older trials, no computer tomography scans were performed to assess response rates, but often clinical examinations and x-ray of the chest were used. Thus, the overall results presented here should be seen as guidance for future large clinical trials and provide an extensive overview of the current evidence of different immunotherapies. Furthermore, besides the immunotherapies reported here, nonspecific immunomodulators, including but not limited to herbal extracts, interleukin or cytokine therapy, can be used to modulate the immune response to fight cervical cancer. However, our search strategies were not able to reliably detect all of these trials. Thus, they were excluded as the risk of missing relevant trials was too high to achieve the standard of a systematic review. Nevertheless, we have supplied an exemplary overview of various immunomodulating agents tested in cervical cancer patients in the Table S1.
To promote the role of immunotherapy in cervical cancer, larger clinical trials are needed, and few are currently ongoing. In particular, therapeutic cancer vaccines have not yet been assessed in large clinical trials, despite the success of prophylactic vaccinations in cervical cancer. However, besides evaluating the efficacy of currently known drugs that have shown promising results in phase I and II trials, new approaches to modify the body's immune response as well as to increase the responsiveness of tumor cells to immunotherapy need to be developed, as response rates to immunotherapy remain low. Promising strategies include the combination of targeted and untargeted immunotherapies as well as increasing the amount of immunotherapeutic target structures on tumors by inhibiting their destruction or potentially inducing their expression.

Conclusions
Immunotherapy in cervical cancer is on the uprise. The first results of high-quality trials on ICIs have led to the approval of pembrolizumab for cervical cancer as a monotherapy and, recently, in combination with CHT and bevacizumab by the FDA. These results have changed the standard of care for patients with persistent, recurrent or metastatic PD-L1-positive cervical cancer. On the other hand, no equivalent immunotherapy option is currently available for PD-L1-negative patients who do not profit from ICIs. Despite still being at the beginning of clinical testing, therapeutic vaccines and ACTT are promising options and have shown some spectacular remissions, even in heavily pretreated patients. However, the overall response rates remain low. Initial investigations demonstrated the potential of combining different immunotherapeutic approaches to increase effectiveness due to synergistic mechanisms of action. As expected, common side effects were immunerelated, and overall, ICIs as well as therapeutic vaccines have proven to be safe and are generally well tolerated, even during combination therapy with RT, CHT or CRT. Similarly, ACTT has not led to treatment-related deaths; however, the preceding non-myeloablative CHT, in particular, causes an increased rate of severe adverse events. Thus, the use of ICIs in fragile, elderly patients should be considered carefully even in clinical trials. All things considered, further clinical trials are needed to verify the effects of immunotherapy as single agents or as combination therapies in larger cohorts.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/cancers14020441/s1, Table S1: Overview of unspecific immunomodulating therapies. Figure S1: Risk of bias assessment of included randomized controlled trialsindividual trial judgements. Figure S2: Risk of bias assessment of included randomized controlled trials-collective risk of bias assessments. Figure S3: Risk of bias assessment of included single-arm cohort trials-individual trial judgements. Figure S4: Risk of bias assessment of included single-arm cohort trials-collective risk of bias assessments. Appendix A: Complete search strategies for Ovid, Web of Science and the Cochrane Library. Funding: This systematic review was conducted independently. No funding was received for this study.