Abstract
Objective
To investigate the efficacy of imiquimod in women with residual or recurrent cervical intraepithelial neoplasia (rrCIN), compared with large loop excision of the transformation zone (LLETZ).
Design
Randomised controlled non-inferiority trial.
Setting
One academic and one regional hospital in the Netherlands.
Population
Thirty-five women with rrCIN were included in the study between May 2016 and May 2021.
Methods
Women were randomised to receive treatment with 5% imiquimod cream (12.5 mg) intravaginally (three times a week for a duration of 16 weeks) or a LLETZ procedure (standard treatment).
Main outcome measures
The primary outcome was reduction to normal cytology at 6 months after starting treatment. Secondary outcomes were clearance of high-risk human papilloma virus (hr-HPV) in both groups and reduction to ≤CIN1 in the imiquimod group. Side effects were monitored.
Results
Treatment success was 33% (6/18) in the imiquimod group versus 100% (16/16) in the LLETZ group (P < 0.001), whereas HPV clearance was 22% (4/18) in the imiquimod group versus 88% (14/16) in the LLETZ group (P < 0.001). After the randomisation of 35 women, the futility of treatment with imiquimod was proven and the trial was prematurely finished. In the follow-up period, three patients remained without additional treatment, whereas all other patients underwent LLETZ, conisation or hysterectomy. In the LLETZ group none of the patients received additional treatment during 2 years of follow-up.
Conclusions
This is the first randomised controlled trial to show that topical imiquimod has a significantly lower success rate in terms of reduction to normal cytology and hr-HPV clearance, compared with LLETZ, in women with rrCIN. Additionally, imiquimod has numerous side effects and after using imiquimod most women with rrCIN still required additional surgical treatment.
1 INTRODUCTION
Cervical intraepithelial neoplasia (CIN) is a precursor of cervical carcinoma caused by human papillomavirus (HPV).1 CIN is classified either as low-grade squamous intraepithelial lesions (L-SIL, CIN 1) or as high-grade squamous intraepithelial lesions (H-SIL, CIN 2–3). As CIN 2 and 3 have a 5% and 12% chance, respectively, of progressing to cervical cancer, both stages are usually treated to prevent progression to invasive cervical cancer.2 In contrast to many other forms of cancer, both cervical carcinoma and CIN often develop at a relatively young age, which presents specific challenges in terms of treatment options.
The gold standard for treatment of H-SIL of the cervix is large loop excision of the transformation zone (LLETZ), which aims to remove the transformation zone. This method can be performed under local anaesthesia with low morbidity, and also provides histological confirmation of disease status.3 Nevertheless, LLETZ has several surgical side effects, of which the increased risk of premature birth in subsequent pregnancies may be the most relevant.4–6 This risk increases with cone depth and number of treatments, and the risk of premature birth can be as high as 13%.7
As residual or recurrent disease after LLETZ is reported in 5%–22% of cases,3, 8–11 this can be especially problematic, with many women of childbearing age being diagnosed with CIN and wishing to avoid risks in subsequent pregnancies. Thus, the quest for a less invasive but equally effective treatment continues. Alternatives have been sought in immune-modulating, antiproliferative and antiviral therapies.12 Of the immune modulators, imiquimod has received the most attention.
Imiquimod is an immune-response modulator that induces the local production of cytokines and infiltration of macrophages and natural killer cells, which leads to a reduction in HPV replication.13 Imiquimod has already proven effective as a first-line treatment of vulvar high-grade squamous intraepithelial lesion (VHSIL), which is also caused by high-risk HPV (hr-HPV).14 Furthermore, imiquimod can achieve a histological regression rate as high as 55% in women with primary high-grade CIN.15 However, treatment with imiquimod is also associated with frequent side effects, which are only acceptable if they are accompanied by a high success rate.16
The TOPIC-2 trial is a randomised controlled non-inferiority trial with two intervention arms, comparing imiquimod with LLETZ in patients with residual or recurrent CIN (rrCIN) lesions. The aim of the study was to determine the efficacy, safety and potential advantages of imiquimod therapy in women with rrCIN.
2 METHODS
2.1 Study design
The TOPIC-2 study was designed as a non-inferiority trial on the efficacy of imiquimod in rrCIN lesions. The study was conducted in the Erasmus MC, Rotterdam, and in the Meander Medical Centre, Amersfoort, the Netherlands. Inclusion took place at the outpatient clinics of the Departments of Gynaecology in the respective hospitals between May 2016 and May 2021. Approval was obtained from the Medical Ethics Committee of the Erasmus Medical Centre (NL53792.078.15/METC-2015-389, 20 October 2015) and the study protocol has been published previously (ClinicalTrials.gov NCT02669459, 27 January 2016). The current article was written according to the Consolidated Standards of Reporting Trials (CONSORT) checklist and guidelines.17
2.2 Patient population
Women aged 18 years or older were assessed for eligibility based on either rrCIN 2–3 or a persistent diagnosis of rrCIN 1, all based on cervical biopsies. Than it’s clear CIN 1 is also residual or recurrent. A persistent CIN 1 lesion was defined as a lesion confirmed twice by biopsy over at least a 6-month interval. Eligible patients required a conisation or LLETZ procedure in their medical history at least 6 months before the diagnosis of the current CIN lesion. Additional inclusion criteria were satisfactory colposcopy (i.e. fully visible transformation zone and fully visible lesion), signed informed consent and an appropriate contraception method for women who were fertile.
Exclusion criteria were PAP4 cytology, adenocarcinoma in situ, a history of (micro)invasive cervical cancer, hypersensitivity to imiquimod, immunodeficiency, pregnancy or lactation and insufficient knowledge of the English or Dutch language. Immunodeficiency was defined as women with human immunodeficiency virus (HIV) or the use of immunosuppressive drugs (for instance, azathioprine or prednisone). All women provided written informed consent before they were included in the study.
All eligible patients were informed that expectant management (in the case of CIN 1–2) or LLETZ (in the case of CIN 2–3) were current standard treatments for high-grade squamous intraepithelial lesions (HSILs). A CONSORT flow diagram is provided (Figure 1).
2.3 Patient involvement
The patient organisation Olijf (for women with gynaecological cancer) was involved in the development of this study. Once the trial has been published, participants will be sent details of the results in a study newsletter suitable for a non-specialist audience.
2.4 Randomisation and masking
Patients who were eligible for the study and agreed to participate were randomly assigned to one of two parallel groups receiving either treatment with self-applied imiquimod or treatment with LLETZ. A computerised randomisation tool (ALEA) was used to avoid selection and allocation bias.
The study was single blinded, with the pathologists evaluating cytological and histological samples blinded with respect to the intervention. A central pathology review of the biopsies and LLETZs was performed upon completion of the study.
2.5 Procedures
Patients in the imiquimod treatment group applied 12.5 mg of imiquimod cream intravaginally with an applicator three times a week for 16 weeks. Patients received instructions and administered the cream themselves before bedtime. The patients were also advised to use an intravaginal shower with an applicator the next morning to remove any leftover cream, followed by an external shower to remove any residue on the vulva. Subjects were asked to refrain from vaginal intercourse during the nights that imiquimod was applied until after the vaginal shower the next morning. To prevent pregnancy, patients were advised to use adequate contraception.
Anti-inflammatory drugs (paracetamol or nonsteroidal anti-inflammatory drugs) were used in case of mild systemic drug-related side effects. If the local or systemic side effects persisted or were severe, patients were advised to reduce the frequency of imiquimod treatment, first to twice weekly and subsequently to once weekly. Imiquimod treatment was discontinued for no more than a week in the event of persistent side effects.
After 10 weeks of treatment a colposcopy was performed to rule out disease progression, and biopsies were only performed when there was a suspicion of invasive disease.
In the standard treatment group patients underwent a LLETZ procedure, preferably within 4 weeks after diagnosis. The excision of macroscopic lesions and the transformation zone was achieved by a monopolar loop electrode, if possible, under local anaesthesia.
Treatment efficacy for both groups was evaluated at the follow-up at 26 weeks. The LLETZ group had follow-up with cervical smears at 6 (26 weeks), 12 and 24 months. The imiquimod group had cervical cytology taken and a colposcopy with diagnostic biopsies at 26 weeks after the start of treatment. The biopsies were performed at the initial CIN lesion site and at any other suspected site, with a minimum of two biopsies performed. In the case of persistent or progressive disease, surgical excision was performed, mostly by LLETZ.
Women who stopped treatment with imiquimod after only a few doses were advised on LLETZ treatment, as this remains the standard treatment for patients with rrCIN.
2.6 Outcomes
The primary study outcome was to evaluate the effectiveness of the vaginal application of imiquimod 5% in the treatment of rrCIN, in comparison with LLETZ. Complete regression was defined as a reduction to normal cervix cytology (PAP1) at 26 weeks after starting treatment in the imiquimod and LLETZ groups. No core outcome set was used, as none is available for this condition.
The secondary outcomes included an evaluation of the effectiveness of the vaginal application of imiquimod 5% in the treatment of rrCIN, measured as a reduction to ≤CIN 1 in histology at 26 weeks, compared with baseline, in the imiquimod group. We also evaluated the effect of treatment on the HPV DNA positivity of CIN lesions, using polymerase chain reaction (PCR) to detect HPV DNA in cervical biopsies/cytology taken at 0 and 26 weeks. The incidence and severity of the side effects of LLETZ and imiquimod therapy were also monitored.
2.7 Statistical methods
The analysis was conducted according to the intention-to-treat principle and the non-inferiority principle, to show that imiquimod is not inferior to an existing treatment (LLETZ). The primary statistical outcome was the difference between the probability of regression in the LLETZ and imiquimod arms. The regression rate of a second LLETZ procedure was estimated as 63% in a retrospective cohort. The estimated regression rate of CIN lesions after treatment with imiquimod was based on the only study reporting a regression rate after 16 weeks of imiquimod therapy for primary CIN, which was 73% in patients with primary CIN lesions.18 As the effect might be lower in patients with rrCIN, we chose a regression rate of 60%. Using a non-inferiority margin of 10% for the difference in proportions and a significance level (alpha) of 5%, a sample size of 174 per group would be required to obtain a power of at least 80%. Allowing for 20% loss to follow-up, the total required sample size would be 433. However, owing to uncertainty surrounding these assumptions, we first performed an interim analysis for the futility of treatment with imiquimod as a pilot study with 35 patients. This interim analysis was based on predictive power, so if the predictive power at this point was below 20%, the envisaged trial would be cancelled for futility of treatment, but if the predictive power at this point was above 20% (indicating non-futility), we would proceed to carry out a national multicentre trial including 433 patients.
The Mann–Whitney U-test was used for continuous variables, and Fisher’s exact test and the chi-square test were used for categorical variables. Differences were considered significant for P < 0.05. No subgroup analyses were performed because of the small number of patients.
3 RESULTS
Between May 2016 and May 2021, a total of 35 women were included in the study. The planned interim analysis showed the futility of treatment with imiquimod, and the trial was terminated. Eighteen women were allocated to imiquimod treatment and 17 women were allocated to LLETZ treatment. The trial CONSORT diagram can be found in Figure 1 and the baseline characteristics are presented in Table 1. In the LLETZ group, one woman failed to meet the inclusion criteria and was excluded. Her biopsy at study entry showed a CIN 1, confirmed upon revision by a gynaecopathologist.TABLE 1. Baseline characteristics of study participants according to treatment group.
| Imiquimod group(n = 18) | LLETZ group(n = 16) | P | |
|---|---|---|---|
| Age (median) | 35.11 (28.7–39.8) | 34.70 (31.8–41.3) | 0.58 |
| BMI (kg/m2) | 24.08 | 24.09 | 0.66 |
| Nulliparity, n (%) | 7 (39) | 5 (31) | 0.73 |
| Smoking | 11 | 8 | 0.73 |
| Sexual contacts in the last year, n (%) | |||
| Single sexual contact | 10 | 13 | 0.07 |
| Multiple sexual contacts | 5 | 1 | |
| No sexual contact | 3 | 1 | |
| Missing | 0 | 2 | |
| Number of LLETZ procedures in history | |||
| 1 | 15 | 13 | 0.54 |
| 2 | 2 | 3 | |
| 3 | 1 | 0 | |
| Highest CIN classification in history | |||
| CIN 1 | 0 | 2 | 0.15 |
| CIN 2 | 4 | 1 | |
| CIN 3 | 13 | 13 | |
| Missing | 1 | 0 | |
| Contraception | |||
| Condoms | 6 | 3 | 0.45 |
| Oral contraception | 6 | 6 | 1.00 |
| Mirena | 1 | 4 | 0.16 |
| Depo Provera | 1 | 1 | 1.00 |
| Other | 4 | 3 | 1.00 |
| Wish to conceive | |||
| Yes | 8 | 4 | 0.48 |
| No | 7 | 9 | |
| Maybe | 3 | 3 | |
| Time between last procedure and study start (median) | 15.97 months | 19.06 months | 0.73 |
| Cytology at start study | |||
| L-SIL | 3 | 4 | 0.68 |
| H-SIL | 15 | 12 | |
| CIN classification at start study | |||
| CIN 2 | 5 | 3 | 0.69 |
| CIN 3 | 13 | 13 | |
| hr-HPV positivity | 100% | 100% | 1.00 |
| Follow-up time in months (median) | 30.5 (15.8–45.5) | 18 (8.25–23.5) | 0.25 |
- Abbreviations: BMI, body mass index; CIN, cervical intraepithelial neoplasia; hr-HPV, high-risk human papillomavirus; LLETZ, large loop excision of the transformation zone; L-SIL/H-SIL, low- or high-grade squamous intraepithelial lesion.
There were no exclusions in the imiquimod group, although one woman immediately opted for LLETZ treatment after randomisation. In the imiquimod group, four women discontinued their treatment after starting: two because of side effects, one of whom had severe abnormalities detected in the lab results; another because of non-compliance with the treatment protocol; and the fourth woman opted for LLETZ treatment after a single dose of imiquimod. Of the 13 women who continued imiquimod treatment, nine elected to lower the frequency or to temporarily pause treatment. The median number of imiquimod applications amongst the 13 women was 45 bags (IQR 33–48 bags). In the LLETZ group, treatment adherence was 100%. All participants were available for follow-up after 26 weeks, although for one case in the imiquimod group histology but no cytology was available. To give a full overview, all individual patient information is presented in Appendix S1.
3.1 Treatment effectiveness
3.1.1 Imiquimod treatment
Using intention-to-treat analysis, the treatment success rate of imiquimod was 33% (6/18) (Table 2). For one woman no cytology was available at 26 weeks, but biopsies showed no CIN lesion. Of the remaining 17 patients, six showed a reduction to normal cytology 26 weeks after starting treatment, including three women who stopped treatment prematurely and were treated with LLETZ. Two women tested HPV-negative and one tested HPV-positive. Of the three women who were treated with imiquimod and showed complete remission based on cytology, two had CIN 3 at biopsy at 26 weeks after starting treatment, one of whom was HPV-positive and one of whom was HPV-negative. The third woman had CIN 1 and was HPV-negative. Per-protocol analysis showed an even larger difference between imiquimod and LLETZ therapy, with a success rate of 23% (3/13) versus 100% (17/17).TABLE 2. Outcome 26 weeks after the start of treatment, by intention-to-treat analysis.
| Imiquimod (n = 18) | LLETZ (n = 16) | P | OR (95% CI) | |
|---|---|---|---|---|
| Cytology | ||||
| Reduction to normal cytology | 6 | 16 | <0.001 | 0 (0.00–0.69) |
| HPV | ||||
| Clearance | 4 | 14 | <0.001 | 0.047 (0.004–0.320) |
| Histology | ||||
| Reduction to normal histology | 3 | NA | – | – |
| Reduction to ≤CIN 1 | 6 | NA | – | – |
| Persistence of abnormal histology | 6 | NA | – | – |
| Progression of histology | 1 | NA | – | – |
- Abbreviations: CIN, cervical intraepithelial neoplasia; HPV, human papillomavirus; LLETZ, large loop excision of the transformation zone.
Eleven women in the imiquimod group had abnormal cytology at 26 weeks, two of whom were treated with LLETZ after four and 12 doses of imiquimod, respectively, and with both discontinuing their imiquimod treatment because of side effects. Both women were HPV-positive at 26 weeks, but no biopsies were available. Of the remaining nine women with abnormal cytology, eight were HPV-positive and the HPV status could not be determined for one woman. Regarding histology, one woman had normal histology, two had CIN 1, two had CIN 2 and three had CIN 3.
After the inclusion of 35 patients, the interim analysis was performed for predictive power. The predictive power was less than 0.1%, which, according to the protocol, demonstrated the futility of treatment with imiquimod.
3.1.2 LLETZ treatment
All women treated initially with LLETZ showed PAP1 at 26 weeks of follow-up. All but one were HPV-negative at 26 weeks (Table 2).
3.1.3 hrHPV clearance
All participants were hrHPV-positive at baseline. At 26 weeks, the hrHPV status was not available for two women in the imiquimod group. The clearance of hrHPV was higher in the LLETZ group compared with the imiquimod group (88%, 14/16, vs 22%, 4/18; P < 0.001) (Table 2). hrHPV clearance in relation to histological outcome was evaluated in the imiquimod group only, where one out of six patients with histological regression of a CIN lesion showed hrHPV clearance (in two patients the HPV status was unknown). Its people of the rrCIN group who had peristance or progression.
3.1.4 Side effects
Sixteen patients reported side effects of imiquimod. Headache was most commonly reported (81%, 13/16), followed by fatigue (75%, 12/16), myalgia (69%, 11/16) and vulvar pruritus/pain (69%, 11/16).
In the imiquimod group, two women discontinued their treatment because of side effects. One of them had severe side effects after starting imiquimod, reporting immediate headaches, nausea and myalgia that showed a minimal response to paracetamol. After lowering the frequency of imiquimod application at 2 weeks, numerous complaints persisted, including loss of appetite and a 3 kg weight loss. After 12 applications the woman stopped using imiquimod and a blood exam showed hypernatremia (132 mmol/L), elevated liver enzymes (alanine aminotransferase, ALT 98 U/L; aspartate transaminase, AST 73 U/L) and leukopenia (2.5 × 109/L). After stopping imiquimod treatment these results spontaneously improved and the complaints diminished.
Eight patients reported on side effects with LLETZ treatment. The most frequently reported side effects were pain, discharge and blood loss (all 63%, 5/8), followed by itching (50%, 4/8) and fever (38%, 3/8). One woman was readmitted to hospital because of bleeding after the procedure.
3.2 Follow-up
3.2.1 Imiquimod group
At 26 weeks, five out of 13 women (38%) immediately received surgical therapy after failed imiquimod treatment, including three with LLETZ treatment, one with cold-knife conisation and one with LLETZ treatment to rule out invasive disease followed by a hysterectomy. At 1 year, another two women underwent additional surgery, in one case a LLETZ procedure and in the other case a hysterectomy, for rrCIN lesions. In the subsequent follow-up, three additional women underwent a LLETZ procedure, leaving only three women who did not undergo additional surgery. One of these patients is still in follow-up because of abnormal cytology and has remained HPV-positive after imiquimod treatment.
Among the five women who were initially allocated to the imiquimod group but switched to LLETZ after randomisation, three showed normal cytology at 26 weeks and two still had normal cytology at later follow-up. The third woman was HPV-positive at 26 weeks and later underwent hysterectomy. The remaining two women showed abnormal cytology at 26 weeks and were HPV-positive, but neither of these women received additional treatment in the follow-up.
3.2.2 LLETZ treatment
One woman in the LLETZ group underwent biopsy for abnormal cytology in the follow-up period, but no one in this group had additional surgical treatment.
4 DISCUSSION
4.1 Main findings
The study was terminated prematurely after a scheduled interim analysis, where the futility of treatment with imiquimod was proven. This suggests that topical imiquimod has a significantly lower success rate in women with rrCIN lesions in comparison with LLETZ treatment. Furthermore, at 26 weeks after the start of treatment, 38% of women who completed treatment with imiquimod still required additional surgical treatment, a figure that rose to 77% during the long-term follow-up. We also found a significantly lower rate of HPV clearance in women treated with imiquimod compared with women treated with LLETZ.
4.2 Strengths and limitations
This study has several strengths. First, it is the only randomised trial to date comparing imiquimod with LLETZ in women with rrCIN. Previous studies were mostly small, retrospective studies.19–21 The study was adequately performed and had a long follow-up period. Although the number of women included might appear low, the number chosen was adequate to prove the futility of treatment with imiquimod.
A major limitation of this study was the slow rate of inclusion. As shown in Figure 1, only 59 patients were assessed for eligibility in 5 years, over a third of whom declined to enter the study and mostly requested a LLETZ procedure. As residual disease has been reported in 5%–27% of cases,8, 22 one might expect more women to meet the inclusion criteria for the TOPIC-2 trial. One possibility is that women were not deemed suitable for the trial by the physician, for instance because of age, visibility of the transformation zone or remaining cervical length. Another possibility is that most women with rrCIN prefer surgical treatment. Moreover, even after counselling relatively few women are interested in imiquimod, possibly because of low expectations of success and the high risk of side effects. Nevertheless, at the outset women were counselled on an estimated success rate of 60%. Interestingly, a study by Koeneman et al. predicted that women would be willing to use imiquimod only if the success rate was at least 72%.16 It is unclear whether this also holds for this specific group of women with rrCIN.
4.3 Interpretation
By achieving a success rate of only 33% based on intention-to-treat analysis, and even 28% based on per-protocol analysis, imiquimod fell far short of our expectations (60%). In our recently published meta-analysis that included new studies, we found a histological regression rate to CIN 1 or below in 55% of women never previously treated.15 A possible explanation for the lower success rate in women with primary CIN, in comparison with the first randomised controlled trial, was the means of application of imiquimod (applicator vs suppository).18, 23 Also, in this study an applicator was used. However, this also perhaps suggests that a 60% rate of regression to normal cytology might be unrealistic in a patient group with rrCIN, as this group is likely to be difficult to treat. Most women of fertile age with CIN 2 are advised on expectant management,24 suggesting that only those with a low chance of spontaneous regression are candidates for further treatment.
Another possible issue is that women in this group might be struggling to clear HPV because of a weak anti-HPV immune reaction. There is contradictory evidence on the effectiveness of imiquimod in patients with other immune disorders, such as HIV.25, 26 Equally, imiquimod might be less effective in patients with underlying immune disease or other causes of a weak immune response.
The success rate of imiquimod in rrCIN could even be worse in a regular treatment setting. Given that 33% was achieved in a research setting, where patients are followed closely, this figure could be even lower with poorer treatment compliance. Conversely, the success rate of LLETZ in our study was much higher than the initial estimate of 63%, and other studies have reported success rates of 59%–61%.20, 21 One possible explanation could be that larger re-excisions are performed in a recurrence or in a trial setting.
In a comparison of imiquimod with LLETZ it is difficult to choose the primary outcome, a problem that has also plagued other studies.23, 27 Ideally, histology would be the primary outcome for both groups, but this would mean an extra biopsy from every woman with a LLETZ, even in the case of normal cytology. Both histology and cytology have interobserver variability.28 Cytology was therefore chosen for its availability in both groups. However, cytology is better at predicting high-grade CIN when combined with HPV testing.29 Taking this standpoint would, however, be very strict, as we are unsure whether the total remission of HPV is achievable after only 6 months of treatment with imiquimod. Ultimately, the goal for many young women with rrCIN is regression of the lesion and the avoidance of further treatment. Regardless of these suggested alternative assessments, none would have resulted in equal success rates for imiquimod treatment and LLETZ in this study.
Based on the success rate found here, women with rrCIN are unlikely to choose imiquimod. Further research should first focus on the aetiology and potential immunological factors in rrCIN. Only with a better understanding of the biology of these lesions can we learn how to treat them effectively, with the least possible impact on women, regarding success rates, a future wish to conceive, the duration of treatment and side effects. Biomarkers may be helpful in achieving this goal.30 For now, LLETZ remains the preferred treatment in rrCIN.
5 CONCLUSION
Overall, treatment with imiquimod performed considerably worse than expected, and treatment with LLETZ performed better than expected, in this randomised controlled study involving women with rrCIN. Also, almost all women treated with imiquimod still needed additional surgical treatment afterwards. In our opinion, LLETZ remains the standard treatment for rrCIN. To determine whether certain women could benefit from imiquimod treatment, we are currently performing a study to identify biomarkers that might contribute to this goal
ARYAN'S BLOG 